Authors: Andreas Tsamados, Luciano Floridi, Mariarosaria Taddeo
The widespread integration of autoregressive-large language models (AR-LLMs), such as ChatGPT, across established applications, like search engines, has introduced critical vulnerabilities with uniquely scalable characteristics. In this commentary, we analyse these vulnerabilities, their dependence on natural language as a vector of attack, and their challenges to cybersecurity best practices. We offer recommendations designed to mitigate these challenges.
Authors: James Baker
Spearheaded by retail traders on the website reddit, the GameStop short squeeze of early 2021 shows that social media embeds information that correlates with market movements. This paper seeks to examine this relationship by using daily frequencies of classified comments and buzzwords as additional factors in a Fama-French three factor model. Comments are classified using an unsupervised clustering method, while past studies have used pretrained models that are not specific to the domains being studied.
Authors: Jinghan Yang, Shuming Ma, Furu Wei
In the era of Large Language Models (LLMs), human-computer interaction has evolved towards natural language, offering unprecedented flexibility. Despite this, LLMs are heavily reliant on well-structured prompts to function efficiently within the realm of In-Context Learning. Vanilla In-Context Learning relies on human-provided contexts, such as labeled examples, explicit instructions, or other guiding mechanisms that shape the model's outputs. To address this challenge, our study presents a universal framework named Automatic In-Context Learning. Upon receiving a user's request, we ask the model to independently generate examples, including labels, instructions, or reasoning pathways. The model then leverages this self-produced context to tackle the given problem. Our approach is universally adaptable and can be implemented in any setting where vanilla In-Context Learning is applicable. We demonstrate that our method yields strong performance across a range of tasks, standing up well when compared to existing methods.
Authors: Yew Ken Chia, Guizhen Chen, Luu Anh Tuan, Soujanya Poria, Lidong Bing
Despite the success of chain of thought in enhancing language model reasoning, the underlying process remains less well understood. Although logically sound reasoning appears inherently crucial for chain of thought, prior studies surprisingly reveal minimal impact when using invalid demonstrations instead. Furthermore, the conventional chain of thought does not inform language models on what mistakes to avoid, which potentially leads to more errors. Hence, inspired by how humans can learn from both positive and negative examples, we propose contrastive chain of thought to enhance language model reasoning. Compared to the conventional chain of thought, our approach provides both valid and invalid reasoning demonstrations, to guide the model to reason step-by-step while reducing reasoning mistakes. To improve generalization, we introduce an automatic method to construct contrastive demonstrations. Our experiments on reasoning benchmarks demonstrate that contrastive chain of thought can serve as a general enhancement of chain-of-thought prompting.
Authors: Fangzhi Xu, Zhiyong Wu, Qiushi Sun, Siyu Ren, Fei Yuan, Shuai Yuan, Qika Lin, Yu Qiao, Jun Liu
Large Language Models (LLMs) have greatly propelled the progress in natural language(NL)-centric tasks based on NL interface. However, the NL form is not enough for world knowledge. Current works focus on this question by injecting specific symbolic knowledge into LLM, which ignore two critical challenges: the interrelations between various symbols and the balance between symbolic-centric and NL-centric capabilities. In this work, we tackle these challenges from both a data and framework perspective and introduce Symbol-LLM series models. First, we collect 34 symbolic tasks, covering ~20 different forms, which are unified to capture symbol interrelations. Then, a two-stage tuning framework succeeds in injecting symbolic knowledge without loss of the generality ability. Extensive experiments on both symbol- and NL-centric tasks demonstrate the balanced and superior performances of Symbol-LLM series models.
Authors: Yuchen Zhou, Emmy Liu, Graham Neubig, Leila Wehbe
Do machines and humans process language in similar ways? A recent line of research has hinted in the affirmative, demonstrating that human brain signals can be effectively predicted using the internal representations of language models (LMs). This is thought to reflect shared computational principles between LMs and human language processing. However, there are also clear differences in how LMs and humans acquire and use language, even if the final task they are performing is the same. Despite this, there is little work exploring systematic differences between human and machine language processing using brain data. To address this question, we examine the differences between LM representations and the human brain's responses to language, specifically by examining a dataset of Magnetoencephalography (MEG) responses to a written narrative. In doing so we identify three phenomena that, in prior work, LMs have been found to not capture well: emotional understanding, figurative language processing, and physical commonsense. By fine-tuning LMs on datasets related to these phenomena, we observe that fine-tuned LMs show improved alignment with human brain responses across these tasks. Our study implies that the observed divergences between LMs and human brains may stem from LMs' inadequate representation of these specific types of knowledge.
Authors: Mohammad Amaan Sayeed, Hanan Aldarmaki
Text word embeddings that encode distributional semantic features work by modeling contextual similarities of frequently occurring words. Acoustic word embeddings, on the other hand, typically encode low-level phonetic similarities. Semantic embeddings for spoken words have been previously explored using similar algorithms to Word2Vec, but the resulting vectors still mainly encoded phonetic rather than semantic features. In this paper, we examine the assumptions and architectures used in previous works and show experimentally how Word2Vec algorithms fail to encode distributional semantics when the input units are acoustically correlated. In addition, previous works relied on the simplifying assumptions of perfect word segmentation and clustering by word type. Given these conditions, a trivial solution identical to text-based embeddings has been overlooked. We follow this simpler path using automatic word type clustering and examine the effects on the resulting embeddings, highlighting the true challenges in this task.
Authors: Tong Liu, Iza Škrjanec, Vera Demberg
Past studies have provided broad support for that words with lower predictability (i.e., higher surprisal) require more time for comprehension by using large language models (LLMs) to simulate humans' cognitive load. In general, these studies have implicitly assumed that the probability scores from LLMs are accurate, ignoring the discrepancies between human cognition and LLMs from this standpoint. Inspired by the concept of probability calibration, we are the first work to focus on the probability distribution for human reading simulation. We propose to use temperature-scaled surprisal, a surprisal calculated by shaped probability, to be the predictor of human reading times. Our results across three corpora consistently revealed that such a surprisal can drastically improve the prediction of reading times. Setting the temperature to be approximately 2.5 across all models and datasets can yield up to an 89% of increase in delta log-likelihood in our setting. We also propose a calibration metric to quantify the possible human-likeness bias. Further analysis was done and provided insights into this phenomenon.
Authors: George Chrysostomou, Zhixue Zhao, Miles Williams, Nikolaos Aletras
Despite their remarkable performance on abstractive summarization, large language models (LLMs) face two significant challenges: their considerable size and tendency to hallucinate. Hallucinations are concerning because they erode the reliability of LLMs and raise safety issues. Pruning is a technique that reduces model size by removing redundant weights to create sparse models that enable more efficient inference. Pruned models yield comparable performance to their counterpart full-sized models, making them ideal alternatives when operating on a limited budget. However, the effect that pruning has upon hallucinations in abstractive summarization with LLMs has yet to be explored. In this paper, we provide an extensive empirical study on the hallucinations produced by pruned models across three standard summarization tasks, two pruning approaches, three instruction-tuned LLMs, and three hallucination evaluation metrics. Surprisingly, we find that pruned LLMs hallucinate less compared to their full-sized counterparts. Our follow-up analysis suggests that pruned models tend to depend more on the source input and less on their parametric knowledge from pre-training for generation. This greater dependency on the source input leads to a higher lexical overlap between generated content and the source input, which can be a reason for the reduction in hallucinations.
Authors: Wenda Xu, Daniel Deutsch, Mara Finkelstein, Juraj Juraska, Biao Zhang, Zhongtao Liu, William Yang Wang, Lei Li, Markus Freitag
Recent improvements in text generation have leveraged human feedback to improve the quality of the generated output. However, human feedback is not always available, especially during inference. In this work, we propose an inference time optimization method FITO to use fine-grained actionable feedback in the form of error type, error location and severity level that are predicted by a learned error pinpoint model for iterative refinement. FITO starts with an initial output, then iteratively incorporates the feedback via a refinement model that generates an improved output conditioned on the feedback. Given the uncertainty of consistent refined samples at iterative steps, we formulate iterative refinement into a local search problem and develop a simulated annealing based algorithm that balances exploration of the search space and optimization for output quality. We conduct experiments on three text generation tasks, including machine translation, long-form question answering (QA) and topical summarization. We observe 0.8 and 0.7 MetricX gain on Chinese-English and English-German translation, 4.5 and 1.8 ROUGE-L gain at long form QA and topic summarization respectively, with a single iteration of refinement. With our simulated annealing algorithm, we see further quality improvements, including up to 1.7 MetricX improvements over the baseline approach.
Authors: Alexandra Chronopoulou, Jonas Pfeiffer, Joshua Maynez, Xinyi Wang, Sebastian Ruder, Priyanka Agrawal
Parameter-efficient fine-tuning (PEFT) using labeled task data can significantly improve the performance of large language models (LLMs) on the downstream task. However, there are 7000 languages in the world and many of these languages lack labeled data for real-world language generation tasks. In this paper, we propose to improve zero-shot cross-lingual transfer by composing language or task specialized parameters. Our method composes language and task PEFT modules via element-wise arithmetic operations to leverage unlabeled data and English labeled data. We extend our approach to cases where labeled data from more languages is available and propose to arithmetically compose PEFT modules trained on languages related to the target. Empirical results on summarization demonstrate that our method is an effective strategy that obtains consistent gains using minimal training of PEFT modules.
Authors: Robert Mahari, Dominik Stammbach, Elliott Ash, Alex `Sandy' Pentland
We present the Legal Passage Retrieval Dataset LePaRD. LePaRD is a massive collection of U.S. federal judicial citations to precedent in context. The dataset aims to facilitate work on legal passage prediction, a challenging practice-oriented legal retrieval and reasoning task. Legal passage prediction seeks to predict relevant passages from precedential court decisions given the context of a legal argument. We extensively evaluate various retrieval approaches on LePaRD, and find that classification appears to work best. However, we note that legal precedent prediction is a difficult task, and there remains significant room for improvement. We hope that by publishing LePaRD, we will encourage others to engage with a legal NLP task that promises to help expand access to justice by reducing the burden associated with legal research. A subset of the LePaRD dataset is freely available and the whole dataset will be released upon publication.
Authors: Sridevi Wagle, Sai Munikoti, Anurag Acharya, Sara Smith, Sameera Horawalavithana
Large language models (LLMs) have shown remarkable achievements in natural language processing tasks, producing high-quality outputs. However, LLMs still exhibit limitations, including the generation of factually incorrect information. In safety-critical applications, it is important to assess the confidence of LLM-generated content to make informed decisions. Retrieval Augmented Language Models (RALMs) is relatively a new area of research in NLP. RALMs offer potential benefits for scientific NLP tasks, as retrieved documents, can serve as evidence to support model-generated content. This inclusion of evidence enhances trustworthiness, as users can verify and explore the retrieved documents to validate model outputs. Quantifying uncertainty in RALM generations further improves trustworthiness, with retrieved text and confidence scores contributing to a comprehensive and reliable model for scientific applications. However, there is limited to no research on UQ for RALMs, particularly in scientific contexts. This study aims to address this gap by conducting a comprehensive evaluation of UQ in RALMs, focusing on scientific tasks. This research investigates how uncertainty scores vary when scientific knowledge is incorporated as pretraining and retrieval data and explores the relationship between uncertainty scores and the accuracy of model-generated outputs. We observe that an existing RALM finetuned with scientific knowledge as the retrieval data tends to be more confident in generating predictions compared to the model pretrained only with scientific knowledge. We also found that RALMs are overconfident in their predictions, making inaccurate predictions more confidently than accurate ones. Scientific knowledge provided either as pretraining or retrieval corpus does not help alleviate this issue. We released our code, data and dashboards at https://github.com/pnnl/EXPERT2.
Authors: Rao Ma, Adian Liusie, Mark J. F. Gales, Kate M. Knill
Text and vision foundation models can perform many tasks in a zero-shot setting, a desirable property that enables these systems to be applied in general and low-resource settings. However, there has been significantly less work on the zero-shot abilities of ASR foundation models, with these systems typically fine-tuned to specific tasks or constrained to applications that match their training criterion and data annotation. In this work we investigate the ability of Whisper and MMS, ASR foundation models trained primarily for speech recognition, to perform zero-shot audio classification. We use simple template-based text prompts at the decoder and use the resulting decoding probabilities to generate zero-shot predictions. Without training the model on extra data or adding any new parameters, we demonstrate that Whisper shows promising zero-shot classification performance on a range of 8 audio-classification datasets, outperforming existing state-of-the-art zero-shot baseline's accuracy by an average of 9%. One important step to unlock the emergent ability is debiasing, where a simple unsupervised reweighting method of the class probabilities yields consistent significant performance gains. We further show that performance increases with model size, implying that as ASR foundation models scale up, they may exhibit improved zero-shot performance.
Authors: Jamie McCusker
While the potential of Open Information Extraction (Open IE) for Knowledge Graph Construction (KGC) may seem promising, we find that the alignment of Open IE extraction results with existing knowledge graphs to be inadequate. The advent of Large Language Models (LLMs), especially the commercially available OpenAI models, have reset expectations for what is possible with deep learning models and have created a new field called prompt engineering. We investigate the use of GPT models and prompt engineering for knowledge graph construction with the Wikidata knowledge graph to address a similar problem to Open IE, which we call Open Knowledge Extraction (OKE) using an approach we call the Linked Open Knowledge Extractor (LOKE, pronounced like "Loki"). We consider the entity linking task essential to construction of real world knowledge graphs. We merge the CaRB benchmark scoring approach with data from the TekGen dataset for the LOKE task. We then show that a well engineered prompt, paired with a naive entity linking approach (which we call LOKE-GPT), outperforms AllenAI's OpenIE 4 implementation on the OKE task, although it over-generates triples compared to the reference set due to overall triple scarcity in the TekGen set. Through an analysis of entity linkability in the CaRB dataset, as well as outputs from OpenIE 4 and LOKE-GPT, we see that LOKE-GPT and the "silver" TekGen triples show that the task is significantly different in content from OIE, if not structure. Through this analysis and a qualitative analysis of sentence extractions via all methods, we found that LOKE-GPT extractions are of high utility for the KGC task and suitable for use in semi-automated extraction settings.
Authors: Swapnil Mane, Suman Kundu, Rajesh Sharma
The rise of social media platforms has led to an increase in cyber-aggressive behavior, encompassing a broad spectrum of hostile behavior, including cyberbullying, online harassment, and the dissemination of offensive and hate speech. These behaviors have been associated with significant societal consequences, ranging from online anonymity to real-world outcomes such as depression, suicidal tendencies, and, in some instances, offline violence. Recognizing the societal risks associated with unchecked aggressive content, this paper delves into the field of Aggression Content Detection and Behavioral Analysis of Aggressive Users, aiming to bridge the gap between disparate studies. In this paper, we analyzed the diversity of definitions and proposed a unified cyber-aggression definition. We examine the comprehensive process of Aggression Content Detection, spanning from dataset creation, feature selection and extraction, and detection algorithm development. Further, we review studies on Behavioral Analysis of Aggression that explore the influencing factors, consequences, and patterns associated with cyber-aggressive behavior. This systematic literature review is a cross-examination of content detection and behavioral analysis in the realm of cyber-aggression. The integrated investigation reveals the effectiveness of incorporating sociological insights into computational techniques for preventing cyber-aggressive behavior. Finally, the paper concludes by identifying research gaps and encouraging further progress in the unified domain of socio-computational aggressive behavior analysis.
Authors: Pritom Saha Akash, Kashob Kumar Roy, Lucian Popa, Kevin Chen-Chuan Chang
Long-form question answering (LFQA) poses a challenge as it involves generating detailed answers in the form of paragraphs, which go beyond simple yes/no responses or short factual answers. While existing QA models excel in questions with concise answers, LFQA requires handling multiple topics and their intricate relationships, demanding comprehensive explanations. Previous attempts at LFQA focused on generating long-form answers by utilizing relevant contexts from a corpus, relying solely on the question itself. However, they overlooked the possibility that the question alone might not provide sufficient information to identify the relevant contexts. Additionally, generating detailed long-form answers often entails aggregating knowledge from diverse sources. To address these limitations, we propose an LFQA model with iterative Planning, Retrieval, and Generation. This iterative process continues until a complete answer is generated for the given question. From an extensive experiment on both an open domain and a technical domain QA dataset, we find that our model outperforms the state-of-the-art models on various textual and factual metrics for the LFQA task.
Authors: Joshua Maher
We introduce a new construction of embeddings of arbitrary recursive data structures into high dimensional vectors. These embeddings provide an interpretable model for the latent state vectors of transformers. We demonstrate that these embeddings can be decoded to the original data structure when the embedding dimension is sufficiently large. This decoding algorithm has a natural implementation as a transformer. We also show that these embedding vectors can be manipulated directly to perform computations on the underlying data without decoding. As an example we present an algorithm that constructs the embedded parse tree of an embedded token sequence using only vector operations in embedding space.
Authors: Jonas Vestergaard Jensen, Mikkel Jordahn, Michael Riis Andersen
In this work, we address the problem of assessing and constructing feedback for early-stage writing automatically using machine learning. Early-stage writing is typically vastly different from conventional writing due to phonetic spelling and lack of proper grammar, punctuation, spacing etc. Consequently, early-stage writing is highly non-trivial to analyze using common linguistic metrics. We propose to use sequence-to-sequence models for "translating" early-stage writing by students into "conventional" writing, which allows the translated text to be analyzed using linguistic metrics. Furthermore, we propose a novel robust likelihood to mitigate the effect of noise in the dataset. We investigate the proposed methods using a set of numerical experiments and demonstrate that the conventional text can be predicted with high accuracy.
Authors: Nalin Kumar, Ondřej Dušek
Linguistic entrainment, or alignment, represents a phenomenon where linguistic patterns employed by conversational participants converge to one another. While alignment has been shown to produce a more natural user experience, most dialogue systems do not have any provisions for it. In this work, we introduce methods for achieving dialogue alignment in a GPT-2-based end-to-end dialogue system through the utilization of shared vocabulary. We experiment with training instance weighting, alignment-specific loss, and additional conditioning to generate responses that align with the user. By comparing different entrainment techniques on the MultiWOZ dataset, we demonstrate that all three approaches produce significantly better-aligned results than the baseline, as confirmed by both automated and manual evaluation metrics.
Authors: Benedikt Ebing, Goran Glavaš
Perfect machine translation (MT) would render cross-lingual transfer (XLT) by means of multilingual language models (LMs) superfluous. Given, on one hand, the large body of work on improving XLT with multilingual LMs and, on the other hand, recent advances in massively multilingual MT, in this work, we systematically evaluate existing and propose new translation-based XLT approaches for transfer to low-resource languages. We show that all translation-based approaches dramatically outperform zero-shot XLT with multilingual LMs, rendering the approach that combines the round-trip translation of the source-language training data with the translation of the target-language test instances the most effective. We next show that one can obtain further empirical gains by adding reliable translations to other high-resource languages to the training data. Moreover, we propose an effective translation-based XLT strategy even for languages not supported by the MT system. Finally, we show that model selection for XLT based on target-language validation data obtained with MT outperforms model selection based on the source-language data. We hope that our findings encourage adoption of more robust translation-based baselines in XLT research.
Authors: James Bernhard
The transformer neural network architecture uses a form of attention in which the dot product of query and key is divided by the square root of the key dimension before applying softmax. This scaling of the dot product is designed to avoid the absolute value of the dot products becoming so large that applying softmax leads to vanishing gradients. In this paper, we propose some alternative scalings, including dividing the dot product instead by the sum of the key lengths before applying softmax. We use simulated keys and queries to show that in many situations this appears to be more effective at avoiding regions where applying softmax leads to vanishing gradients.
Authors: Leonardo Ranaldi, Giulia Pucci
Large Language Models (LLMs) have been demonstrating the ability to solve complex tasks by delivering answers that are positively evaluated by humans due in part to the intensive use of human feedback that refines responses. However, the suggestibility transmitted through human feedback increases the inclination to produce responses that correspond to the user's beliefs or misleading prompts as opposed to true facts, a behaviour known as sycophancy. This phenomenon decreases the bias, robustness, and, consequently, their reliability.
In this paper, we shed light on the suggestibility of LLMs to sycophantic behaviour, demonstrating these tendencies via human-influenced prompts over different tasks. Our investigation reveals that LLMs show sycophantic tendencies when responding to queries involving subjective opinions and statements that should elicit a contrary response based on facts, demonstrating a lack of robustness.
Authors: Yuekun Yao, Alexander Koller
The ability to predict an NLP model's accuracy on unseen, potentially out-of-distribution data is a prerequisite for trustworthiness. We present a novel model that establishes upper and lower bounds on the accuracy, without requiring gold labels for the unseen data. We achieve this by training a discriminator which predicts whether the output of a given sequence-to-sequence model is correct or not. We show across a variety of tagging, parsing, and semantic parsing tasks that the gold accuracy is reliably between the predicted upper and lower bounds, and that these bounds are remarkably close together.
Authors: Yueqing Liang, Lu Cheng, Ali Payani, Kai Shu
This work investigates the potential of undermining both fairness and detection performance in abusive language detection. In a dynamic and complex digital world, it is crucial to investigate the vulnerabilities of these detection models to adversarial fairness attacks to improve their fairness robustness. We propose a simple yet effective framework FABLE that leverages backdoor attacks as they allow targeted control over the fairness and detection performance. FABLE explores three types of trigger designs (i.e., rare, artificial, and natural triggers) and novel sampling strategies. Specifically, the adversary can inject triggers into samples in the minority group with the favored outcome (i.e., ``non-abusive'') and flip their labels to the unfavored outcome, i.e., ``abusive''. Experiments on benchmark datasets demonstrate the effectiveness of FABLE attacking fairness and utility in abusive language detection.
Authors: William Brandon, Aniruddha Nrusimha, Kevin Qian, Zachary Ankner, Tian Jin, Zhiye Song, Jonathan Ragan-Kelley
To help address the growing demand for ever-longer sequence lengths in transformer models, Liu et al. recently proposed Ring Attention, an exact attention algorithm capable of overcoming per-device memory bottle- necks by distributing self-attention across multiple devices. In this paper, we study the performance characteristics of Ring Attention in the important special case of causal transformer models, and identify a key workload imbal- ance due to triangular structure of causal attention computations. We propose a simple extension to Ring Attention, which we call Striped Attention to fix this imbalance. Instead of devices having contiguous subsequences, each device has a subset of tokens distributed uniformly throughout the sequence, which we demonstrate leads to more even workloads. In experiments running Striped Attention on A100 GPUs and TPUv4s, we are able to achieve up to 1.45x end-to-end throughput improvements over the original Ring Attention algorithm on causal transformer training at a sequence length of 256k. Furthermore, on 16 TPUv4 chips, we were able to achieve 1.65x speedups at sequence lengths of 786k. We release the code for our experiments as open source
Authors: Haoran Wang, Kai Shu
To ensure AI safety, instruction-tuned Large Language Models (LLMs) are specifically trained to ensure alignment, which refers to making models behave in accordance with human intentions. While these models have demonstrated commendable results on various safety benchmarks, the vulnerability of their safety alignment has not been extensively studied. This is particularly troubling given the potential harm that LLMs can inflict. Existing attack methods on LLMs often rely on poisoned training data or the injection of malicious prompts. These approaches compromise the stealthiness and generalizability of the attacks, making them susceptible to detection. Additionally, these models often demand substantial computational resources for implementation, making them less practical for real-world applications. In this work, we introduce a novel attack framework, called Backdoor Activation Attack, which injects trojan steering vectors into the activation layers of LLMs. These malicious steering vectors can be triggered at inference time to steer the models toward attacker-desired behaviors by manipulating their activations. In particular, the steering vectors are generated by taking the difference between benign and malicious activations. Then, the most effective steering vector is selected and added to the forward passes of the LLMs. Our experiment results on four primary alignment tasks show that our proposed method is highly effective and adds little or no overhead to attack efficiency. Additionally, we discuss potential countermeasures against such activation attacks. Our code and data are available at https://email-haoran-for-link. Warning: this paper contains content that can be offensive or upsetting.
Authors: Kyle Seelman, Mozhi Zhang, Jordan Boyd-Graber
Topic models help users understand large document collections; however, topic models do not always find the ``right'' topics. While classical probabilistic and anchor-based topic models have interactive variants to guide models toward better topics, such interactions are not available for neural topic models such as the embedded topic model (\abr{etm}). We correct this lacuna by adding an intuitive interaction to neural topic models: users can label a topic with a word, and topics are updated so that the topic words are close to the label. This allows a user to refine topics based on their information need. While, interactivity is intuitive for \abr{etm}, we extend this framework to work with other neural topic models as well. We develop an interactive interface which allows users to interact and relabel topic models as they see fit. We evaluate our method through a human study, where users can relabel topics to find relevant documents. Using our method, user labeling improves document rank scores, helping to find more relevant documents to a given query when compared to no user labeling.
Authors: Brooklyn Sheppard, Anna Richter, Allison Cohen, Elizabeth Allyn Smith, Tamara Kneese, Carolyne Pelletier, Ioana Baldini, Yue Dong
Using novel approaches to dataset development, the Biasly dataset captures the nuance and subtlety of misogyny in ways that are unique within the literature. Built in collaboration with multi-disciplinary experts and annotators themselves, the dataset contains annotations of movie subtitles, capturing colloquial expressions of misogyny in North American film. The dataset can be used for a range of NLP tasks, including classification, severity score regression, and text generation for rewrites. In this paper, we discuss the methodology used, analyze the annotations obtained, and provide baselines using common NLP algorithms in the context of misogyny detection and mitigation. We hope this work will promote AI for social good in NLP for bias detection, explanation, and removal.
Authors: Lingbo Mo, Boshi Wang, Muhao Chen, Huan Sun
The rapid progress in open-source Large Language Models (LLMs) is significantly driving AI development forward. However, there is still a limited understanding of their trustworthiness. Deploying these models at scale without sufficient trustworthiness can pose significant risks, highlighting the need to uncover these issues promptly. In this work, we conduct an assessment of open-source LLMs on trustworthiness, scrutinizing them across eight different aspects including toxicity, stereotypes, ethics, hallucination, fairness, sycophancy, privacy, and robustness against adversarial demonstrations. We propose an enhanced Chain of Utterances-based (CoU) prompting strategy by incorporating meticulously crafted malicious demonstrations for trustworthiness attack. Our extensive experiments encompass recent and representative series of open-source LLMs, including Vicuna, MPT, Falcon, Mistral, and Llama 2. The empirical outcomes underscore the efficacy of our attack strategy across diverse aspects. More interestingly, our result analysis reveals that models with superior performance in general NLP tasks do not always have greater trustworthiness; in fact, larger models can be more vulnerable to attacks. Additionally, models that have undergone instruction tuning, focusing on instruction following, tend to be more susceptible, although fine-tuning LLMs for safety alignment proves effective in mitigating adversarial trustworthiness attacks.
Authors: Prafulla Kumar Choubey, Alexander R. Fabbri, Caiming Xiong, Chien-Sheng Wu
Ideal summarization models should generalize to novel summary-worthy content without remembering reference training summaries by rote. However, a single average performance score on the entire test set is inadequate in determining such model competencies. We propose a fine-grained evaluation protocol by partitioning a test set based on the lexical similarity of reference test summaries with training summaries. We observe up to a 5x (1.2x) difference in ROUGE-2 (entity recall) scores between the subsets with the lowest and highest similarity. Next, we show that such training repetitions also make a model vulnerable to rote learning, reproducing data artifacts such as factual errors, especially when reference test summaries are lexically close to training summaries. Consequently, we propose to limit lexical repetitions in training summaries during both supervised fine-tuning and likelihood calibration stages to improve the performance on novel test cases while retaining average performance. Our automatic and human evaluations on novel test subsets and recent news articles show that limiting lexical repetitions in training summaries can prevent rote learning and improve generalization.
Authors: Yifu Qiu, Varun Embar, Shay B. Cohen, Benjamin Han
Neural knowledge-to-text generation models often struggle to faithfully generate descriptions for the input facts: they may produce hallucinations that contradict the given facts, or describe facts not present in the input. To reduce hallucinations, we propose a novel decoding method, TWEAK (Think While Effectively Articulating Knowledge). TWEAK treats the generated sequences at each decoding step and its future sequences as hypotheses, and ranks each generation candidate based on how well their corresponding hypotheses support the input facts using a Hypothesis Verification Model (HVM). We first demonstrate the effectiveness of TWEAK by using a Natural Language Inference (NLI) model as the HVM and report improved faithfulness with minimal impact on the quality. We then replace the NLI model with our task-specific HVM trained with a first-of-a-kind dataset, FATE (Fact-Aligned Textual Entailment), which pairs input facts with their faithful and hallucinated descriptions with the hallucinated spans marked. The new HVM improves the faithfulness and the quality further and runs faster. Overall the best TWEAK variants improve on average 2.22/7.17 points on faithfulness measured by FactKB over WebNLG and TekGen/GenWiki, respectively, with only 0.14/0.32 points degradation on quality measured by BERTScore over the same datasets. Since TWEAK is a decoding-only approach, it can be integrated with any neural generative model without retraining.
Authors: Michael J.Q. Zhang, Eunsol Choi
Resolving ambiguities through interaction is a hallmark of natural language, and modeling this behavior is a core challenge in crafting AI assistants. In this work, we study such behavior in LMs by proposing a task-agnostic framework for resolving ambiguity by asking users clarifying questions. Our framework breaks down this objective into three subtasks: (1) determining when clarification is needed, (2) determining what clarifying question to ask, and (3) responding accurately with the new information gathered through clarification. We evaluate systems across three NLP applications: question answering, machine translation and natural language inference. For the first subtask, we present a novel uncertainty estimation approach, intent-sim, that determines the utility of querying for clarification by estimating the entropy over user intents. Our method consistently outperforms existing uncertainty estimation approaches at identifying predictions that will benefit from clarification. When only allowed to ask for clarification on 10% of examples, our system is able to double the performance gains over randomly selecting examples to clarify. Furthermore, we find that intent-sim is robust, demonstrating improvements across a wide range of NLP tasks and LMs. Together, our work lays foundation for studying clarifying interactions with LMs.
Authors: Ninareh Mehrabi, Palash Goyal, Anil Ramakrishna, Jwala Dhamala, Shalini Ghosh, Richard Zemel, Kai-Wei Chang, Aram Galstyan, Rahul Gupta
With the recent surge of language models in different applications, attention to safety and robustness of these models has gained significant importance. Here we introduce a joint framework in which we simultaneously probe and improve the robustness of a black-box target model via adversarial prompting and belief augmentation using iterative feedback loops. This framework utilizes an automated red teaming approach to probe the target model, along with a belief augmenter to generate instructions for the target model to improve its robustness to those adversarial probes. Importantly, the adversarial model and the belief generator leverage the feedback from past interactions to improve the effectiveness of the adversarial prompts and beliefs, respectively. In our experiments, we demonstrate that such a framework can reduce toxic content generation both in dynamic cases where an adversary directly interacts with a target model and static cases where we use a static benchmark dataset to evaluate our model.
Authors: Jon Saad-Falcon, Omar Khattab, Christopher Potts, Matei Zaharia
Evaluating retrieval-augmented generation (RAG) systems traditionally relies on hand annotations for input queries, passages to retrieve, and responses to generate. We introduce ARES, an Automated RAG Evaluation System, for evaluating RAG systems along the dimensions of context relevance, answer faithfulness, and answer relevance. Using synthetic training data, ARES finetunes lightweight LM judges to assess the quality of individual RAG components. To mitigate potential prediction errors, ARES utilizes a small set of human-annotated datapoints for prediction-powered inference (PPI). Across six different knowledge-intensive tasks in KILT and SuperGLUE, ARES accurately evaluates RAG systems while using a few hundred human annotations during evaluation. Furthermore, ARES judges remain effective across domain shifts, proving accurate even after changing the type of queries and/or documents used in the evaluated RAG systems. We make our datasets and code for replication and deployment available at https://github.com/stanford-futuredata/ARES.
Authors: Nicholas Lourie, Kyunghyun Cho, He He
The choice of hyperparameters greatly impacts performance in natural language processing. Often, it is hard to tell if a method is better than another or just better tuned. Tuning curves fix this ambiguity by accounting for tuning effort. Specifically, they plot validation performance as a function of the number of hyperparameter choices tried so far. While several estimators exist for these curves, it is common to use point estimates, which we show fail silently and give contradictory results when given too little data.
Beyond point estimates, confidence bands are necessary to rigorously establish the relationship between different approaches. We present the first method to construct valid confidence bands for tuning curves. The bands are exact, simultaneous, and distribution-free, thus they provide a robust basis for comparing methods.
Empirical analysis shows that while bootstrap confidence bands, which serve as a baseline, fail to approximate their target confidence, ours achieve it exactly. We validate our design with ablations, analyze the effect of sample size, and provide guidance on comparing models with our method. To promote confident comparisons in future work, we release a library implementing the method at https://github.com/nalourie/opda .
Authors: Yue Guo, Joseph Chee Chang, Maria Antoniak, Erin Bransom, Trevor Cohen, Lucy Lu Wang, Tal August
Scientific jargon can impede researchers when they read materials from other domains. Current methods of jargon identification mainly use corpus-level familiarity indicators (e.g., Simple Wikipedia represents plain language). However, researchers' familiarity of a term can vary greatly based on their own background. We collect a dataset of over 10K term familiarity annotations from 11 computer science researchers for terms drawn from 100 paper abstracts. Analysis of this data reveals that jargon familiarity and information needs vary widely across annotators, even within the same sub-domain (e.g., NLP). We investigate features representing individual, sub-domain, and domain knowledge to predict individual jargon familiarity. We compare supervised and prompt-based approaches, finding that prompt-based methods including personal publications yields the highest accuracy, though zero-shot prompting provides a strong baseline. This research offers insight into features and methods to integrate personal data into scientific jargon identification.
Authors: Evgeniia Razumovskaia, Goran Glavaš, Anna Korhonen, Ivan Vulić
Task-oriented dialogue (ToD) systems help users execute well-defined tasks across a variety of domains (e.g., $\textit{flight booking}$ or $\textit{food ordering}$), with their Natural Language Understanding (NLU) components being dedicated to the analysis of user utterances, predicting users' intents ($\textit{Intent Detection}$, ID) and extracting values for informational slots ($\textit{Value Extraction}$, VE). In most domains, labelled NLU data is scarce, making sample-efficient learning -- enabled with effective transfer paradigms -- paramount. In this work, we introduce SQATIN, a new framework for dialog NLU based on (i) instruction tuning and (ii) question-answering-based formulation of ID and VE tasks. According to the evaluation on established NLU benchmarks, SQATIN sets the new state of the art in dialogue NLU, substantially surpassing the performance of current models based on standard fine-tuning objectives in both in-domain training and cross-domain transfer. SQATIN yields particularly large performance gains in cross-domain transfer, owing to the fact that our QA-based instruction tuning leverages similarities between natural language descriptions of classes (i.e., slots and intents) across domains.
Authors: Yuxin Pei, Pushkar Bhuse, Zhengzhong Liu, Eric Xing
Interpolation-based Data Augmentation (DA) methods (Mixup) linearly interpolate the inputs and labels of two or more training examples. Mixup has more recently been adapted to the field of Natural Language Processing (NLP), mainly for sequence labeling tasks. However, such a simple adoption yields mixed or unstable improvements over the baseline models. We argue that the direct-adoption methods do not account for structures in NLP tasks. To this end, we propose SegMix, a collection of interpolation-based DA algorithms that can adapt to task-specific structures. SegMix poses fewer constraints on data structures, is robust to various hyperparameter settings, applies to more task settings, and adds little computational overhead. In the algorithm's core, we apply interpolation methods on task-specific meaningful segments, in contrast to applying them on sequences as in prior work. We find SegMix to be a flexible framework that combines rule-based DA methods with interpolation-based methods, creating interesting mixtures of DA techniques. We show that SegMix consistently improves performance over strong baseline models in Named Entity Recognition (NER) and Relation Extraction (RE) tasks, especially under data-scarce settings. Furthermore, this method is easy to implement and adds negligible training overhead.
Authors: Yash Kumar Lal, Li Zhang, Faeze Brahman, Bodhisattwa Prasad Majumder, Peter Clark, Niket Tandon
How-to procedures, such as how to plant a garden, are ubiquitous. But one size does not fit all - humans often need to customize these procedural plans according to their specific needs, e.g., planting a garden without pesticides. While LLMs can fluently generate generic procedures, we present the first study on how well LLMs can customize open-domain procedures. We introduce CustomPlans, a probe dataset of customization hints that encodes diverse user needs for open-domain How-to procedures. Using LLMs as CustomizationAgent and ExecutionAgent in different settings, we establish their abilities to perform open-domain procedure customization. Human evaluation shows that using these agents in a Sequential setting is the best, but they are good enough only ~51% of the time. Error analysis shows that LLMs do not sufficiently address user customization needs in their generated procedures.
Authors: Quinn Patwardhan, Grace Hui Yang
This paper contains what the Georgetown InfoSense group has done in regard to solving the challenges presented by TREC iKAT 2023. Our submitted runs outperform the median runs by a significant margin, exhibiting superior performance in nDCG across various cut numbers and in overall success rate. Our approach uses a Generate-Retrieve-Generate method, which we've found to greatly outpace Retrieve-Then-Generate approaches for the purposes of iKAT. Our solution involves the use of Large Language Models (LLMs) for initial answers, answer grounding by BM25, passage quality filtering by logistic regression, and answer generation by LLMs again. We leverage several purpose-built Language Models, including BERT, Chat-based, and text-to-transfer-based models, for text understanding, classification, generation, and summarization. The official results of the TREC evaluation contradict our initial self-evaluation, which may suggest that a decrease in the reliance on our retrieval and classification methods is better. Nonetheless, our findings suggest that the sequence of involving these different components matters, where we see an essentiality of using LLMs before using search engines.
Authors: Yixiao Song, Kalpesh Krishna, Rajesh Bhatt, Kevin Gimpel, Mohit Iyyer
Grammatical error correction tools are effective at correcting grammatical errors in users' input sentences but do not provide users with \textit{natural language} explanations about their errors. Such explanations are essential for helping users learn the language by gaining a deeper understanding of its grammatical rules (DeKeyser, 2003; Ellis et al., 2006). To address this gap, we propose the task of grammar error explanation, where a system needs to provide one-sentence explanations for each grammatical error in a pair of erroneous and corrected sentences. We analyze the capability of GPT-4 in grammar error explanation, and find that it only produces explanations for 60.2% of the errors using one-shot prompting. To improve upon this performance, we develop a two-step pipeline that leverages fine-tuned and prompted large language models to perform structured atomic token edit extraction, followed by prompting GPT-4 to generate explanations. We evaluate our pipeline on German and Chinese grammar error correction data sampled from language learners with a wide range of proficiency levels. Human evaluation reveals that our pipeline produces 93.9% and 98.0% correct explanations for German and Chinese data, respectively. To encourage further research in this area, we will open-source our data and code.
Authors: Ben Bogin, Shivanshu Gupta, Peter Clark, Ashish Sabharwal
In-context learning (ICL) is an appealing approach for semantic parsing due to its few-shot nature and improved generalization. However, learning to parse to rare domain-specific languages (DSLs) from just a few demonstrations is challenging, limiting the performance of even the most capable LLMs. In this work, we improve the effectiveness of ICL for semantic parsing by (1) using general-purpose programming languages such as Python instead of DSLs, and (2) augmenting prompts with a structured domain description that includes, e.g., the available classes and functions. We show that both these changes significantly improve accuracy across three popular datasets. Combined, they lead to dramatic improvements (e.g. 7.9% to 66.5% on SMCalFlow compositional split), nearly closing the performance gap between easier i.i.d.\ and harder compositional splits when used with a strong model, and reducing the need for a large number of demonstrations. We find that the resemblance of the target parse language to general-purpose code is a more important factor than the language's popularity in pre-training corpora. Our findings provide an improved methodology for building semantic parsers in the modern context of ICL with LLMs.
Authors: Haoyi Qiu, Kung-Hsiang Huang, Jingnong Qu, Nanyun Peng
Ensuring factual consistency is crucial in various natural language processing tasks, particularly in abstractive summarization, where preserving the integrity of information is paramount. Prior entailment-based approaches often generate factually inconsistent summaries and then train a classifier on the generated data. However, summaries produced by these approaches are either of low coherence or lack error-type coverage. To address these issues, we propose AMRFact, a novel framework that generates factually inconsistent summaries using Abstract Meaning Representation (AMR). Our approach parses factually correct summaries into AMR graphs and injects controlled factual inconsistencies to create negative examples, allowing for coherent factually inconsistent summaries to be generated with high error-type coverage. Additionally, we present a data selection module NegFilter based on natural language inference and BARTScore to ensure the quality of the generated negative samples. Experimental results demonstrate that our approach significantly outperforms previous systems on the AggreFact-SOTA dataset, showcasing its efficacy in assessing factuality in abstractive summarization.
Authors: Zhilin Wang, Yi Dong, Jiaqi Zeng, Virginia Adams, Makesh Narsimhan Sreedhar, Daniel Egert, Olivier Delalleau, Jane Polak Scowcroft, Neel Kant, Aidan Swope, Oleksii Kuchaiev
Existing open-source helpfulness preference datasets do not specify what makes some responses more helpful and others less so. Models trained on these datasets can incidentally learn to model dataset artifacts (e.g. preferring longer but unhelpful responses only due to their length). To alleviate this problem, we collect HelpSteer, a multi-attribute helpfulness dataset annotated for the various aspects that make responses helpful. Specifically, our 37k-sample dataset has annotations for correctness, coherence, complexity, and verbosity in addition to overall helpfulness of responses. Training Llama 2 70B using the HelpSteer dataset with SteerLM technique produces a model that scores 7.54 on MT Bench, which is currently the highest score for open models that do not require training data from more powerful models (e.g. GPT4). We release this dataset with CC-BY-4.0 license at https://huggingface.co/datasets/nvidia/HelpSteer
Authors: Xi Ye, Ruoxi Sun, Sercan Ö. Arik, Tomas Pfister
Large language models (LLMs) have achieved remarkable advancements in natural language understanding, generation, and manipulation of text-based data. However, one major issue towards their widespread deployment in the real world is that they can generate "hallucinated" answers that are not factual. Towards this end, this paper focuses on improving grounding from a holistic perspective with a novel framework, AGREE, Adaptation of LLMs for GRounding EnhancEment. We start with the design of an iterative test-time adaptation (TTA) capability that takes into account the support information generated in self-grounded responses. To effectively enable this capability, we tune LLMs to ground the claims in their responses to retrieved documents by providing citations. This tuning on top of the pre-trained LLMs requires a small amount of data that needs to be constructed in a particular way to learn the grounding information, for which we introduce a data construction method. Our results show that the tuning-based AGREE framework generates better grounded responses with more accurate citations compared to prompting-based approaches.
Authors: Yao Dou, Isadora Krsek, Tarek Naous, Anubha Kabra, Sauvik Das, Alan Ritter, Wei Xu
Self-disclosure, while being common and rewarding in social media interaction, also poses privacy risks. In this paper, we take the initiative to protect the user-side privacy associated with online self-disclosure through identification and abstraction. We develop a taxonomy of 19 self-disclosure categories, and curate a large corpus consisting of 4.8K annotated disclosure spans. We then fine-tune a language model for identification, achieving over 75% in Token F$_1$. We further conduct a HCI user study, with 82\% of participants viewing the model positively, highlighting its real world applicability. Motivated by the user feedback, we introduce the task of self-disclosure abstraction. We experiment with both one-span abstraction and three-span abstraction settings, and explore multiple fine-tuning strategies. Our best model can generate diverse abstractions that moderately reduce privacy risks while maintaining high utility according to human evaluation.
Authors: Neha Srikanth, Rupak Sarkar, Rachel Rudinger, Jordan Boyd-Graber
Questions posed by information-seeking users often contain implicit false or potentially harmful assumptions. In a high-risk domain such as maternal and infant health, a question-answering system must recognize these pragmatic constraints and go beyond simply answering user questions, examining them in context to respond helpfully. To achieve this, we study pragmatic inferences made when mothers ask questions about pregnancy and infant care. Some of the inferences in these questions evade detection by existing methods, risking the possibility of QA systems failing to address them which can have dangerous health and policy implications. We explore the viability of detecting inferences from questions using large language models and illustrate that informing existing QA pipelines with pragmatic inferences produces responses that can mitigate the propagation of harmful beliefs.
Authors: Qingyuan Li, Ran Meng, Yiduo Li, Bo Zhang, Liang Li, Yifan Lu, Xiangxiang Chu, Yerui Sun, Yuchen Xie
The large language model era urges faster and less costly inference. Prior model compression works on LLMs tend to undertake a software-centric approach primarily focused on the simulated quantization performance. By neglecting the feasibility of deployment, these approaches are typically disabled in real practice. They used to drastically push down the quantization bit range for a reduced computation which might not be supported by the mainstream hardware, or involve sophisticated algorithms that introduce extra computation or memory access overhead. We argue that pursuing a hardware-centric approach in the construction of quantization algorithms is crucial. In this regard, we are driven to build our compression method on top of hardware awareness, eliminating impractical algorithm choices while maximizing the benefit of hardware acceleration. Our method, OdysseyLLM, comes with a novel W4A8 kernel implementation called FastGEMM and a combined recipe of quantization strategies. Extensive experiments manifest the superiority of our W4A8 method which brings the actual speed boosting up to \textbf{4$\times$} compared to Hugging Face FP16 inference and \textbf{2.23$\times$} vs. the state-of-the-art inference engine TensorRT-LLM in FP16, and \textbf{1.45$\times$} vs. TensorRT-LLM in INT8, yet without substantially harming the performance.
Authors: Tariq Alhindi, Smaranda Muresan, Preslav Nakov
Recognizing fallacies is crucial for ensuring the quality and validity of arguments across various domains. However, computational fallacy recognition faces challenges due to the diverse genres, domains, and types of fallacies found in datasets. This leads to a highly multiclass, and even multi-label, setup with substantial class imbalance. In this study, we aim to enhance existing models for fallacy recognition by incorporating additional context and by leveraging large language models to generate synthetic data, thus increasing the representation of the infrequent classes. We experiment with GPT3.5 to generate synthetic examples and we examine the impact of prompt settings for this. Moreover, we explore zero-shot and few-shot scenarios to evaluate the effectiveness of using the generated examples for training smaller models within a unified fallacy recognition framework. Furthermore, we analyze the overlap between the synthetic data and existing fallacy datasets. Finally, we investigate the usefulness of providing supplementary context for detecting fallacy types that need such context, e.g., diversion fallacies. Our evaluation results demonstrate consistent improvements across fallacy types, datasets, and generators.
Authors: Chaitanya Malaviya, Subin Lee, Dan Roth, Mark Yatskar
Eliciting feedback from end users of NLP models can be beneficial for improving models. However, how should we present model responses to users so they are most amenable to be corrected from user feedback? Further, what properties do users value to understand and trust responses? We answer these questions by analyzing the effect of rationales generated by QA models to support their answers. We specifically consider decomposed question-answering models that first extract an intermediate rationale based on a context and a question and then use solely this rationale to answer the question. A rationale outlines the approach followed by the model to answer the question. Our work considers various formats of these rationales that vary according to well-defined properties of interest. We sample these rationales from large language models using few-shot prompting for two reading comprehension datasets, and then perform two user studies. In the first one, we present users with incorrect answers and corresponding rationales of various formats and ask them to provide natural language feedback to revise the rationale. We then measure the effectiveness of this feedback in patching these rationales through in-context learning. The second study evaluates how well different rationale formats enable users to understand and trust model answers, when they are correct. We find that rationale formats significantly affect how easy it is (1) for users to give feedback for rationales, and (2) for models to subsequently execute this feedback. In addition to influencing critiquablity, certain formats significantly enhance user reported understanding and trust of model outputs.
Authors: Gaurav Sahu, Olga Vechtomova, Issam H. Laradji
This work tackles the task of extractive text summarization in a limited labeled data scenario using a semi-supervised approach. Specifically, we propose a prompt-based pseudolabel selection strategy using GPT-4. We evaluate our method on three text summarization datasets: TweetSumm, WikiHow, and ArXiv/PubMed. Our experiments show that by using an LLM to evaluate and generate pseudolabels, we can improve the ROUGE-1 by 10-20\% on the different datasets, which is akin to enhancing pretrained models. We also show that such a method needs a smaller pool of unlabeled examples to perform better.
Authors: Kuan-Hao Huang, I-Hung Hsu, Tanmay Parekh, Zhiyu Xie, Zixuan Zhang, Premkumar Natarajan, Kai-Wei Chang, Nanyun Peng, Heng Ji
Event extraction has attracted much attention in recent years due to its potential for many applications. However, recent studies observe some evaluation challenges, suggesting that reported scores might not reflect the true performance. In this work, we first identify and discuss these evaluation challenges, including the unfair comparisons resulting from different assumptions about data or different data preprocessing steps, the incompleteness of the current evaluation framework leading to potential dataset bias or data split bias, and low reproducibility of prior studies. To address these challenges, we propose TextEE, a standardized, fair, and reproducible benchmark for event extraction. TextEE contains standardized data preprocessing scripts and splits for more than ten datasets across different domains. In addition, we aggregate and re-implement over ten event extraction approaches published in recent years and conduct a comprehensive reevaluation. Finally, we explore the capability of large language models in event extraction and discuss some future challenges. We expect TextEE will serve as a reliable benchmark for event extraction, facilitating future research in the field.
Authors: Mihir Parmar, Aakanksha Naik, Himanshu Gupta, Disha Agrawal, Chitta Baral
Many large language models (LLMs) for medicine have largely been evaluated on short texts, and their ability to handle longer sequences such as a complete electronic health record (EHR) has not been systematically explored. Assessing these models on long sequences is crucial since prior work in the general domain has demonstrated performance degradation of LLMs on longer texts. Motivated by this, we introduce LongBoX, a collection of seven medical datasets in text-to-text format, designed to investigate model performance on long sequences. Preliminary experiments reveal that both medical LLMs (e.g., BioGPT) and strong general domain LLMs (e.g., FLAN-T5) struggle on this benchmark. We further evaluate two techniques designed for long-sequence handling: (i) local-global attention, and (ii) Fusion-in-Decoder (FiD). Our results demonstrate mixed results with long-sequence handling - while scores on some datasets increase, there is substantial room for improvement. We hope that LongBoX facilitates the development of more effective long-sequence techniques for the medical domain. Data and source code are available at https://github.com/Mihir3009/LongBoX.
Authors: Yao Lu, Jiayi Wang, Sebastian Riedel, Pontus Stenetorp
Using the generative nature of a language model to generate task-relevant separators has shown competitive results compared to human-curated prompts like "TL;DR". We demonstrate that even randomly chosen tokens from the vocabulary as separators can achieve near-state-of-the-art performance. We analyse this phenomenon in detail using three different random generation strategies, establishing that the language space is rich with potential good separators, regardless of the underlying language model size. These observations challenge the common assumption that an effective prompt should be human-readable or task-relevant. Experimental results show that using random separators leads to an average 16% relative improvement across nine text classification tasks on seven language models, compared to human-curated separators, and is on par with automatic prompt searching methods.
Authors: Chen Zhang
AI agents excel in executing predefined tasks, but the dynamic management of work state information during task execution remains an underexplored area. We propose a work state-centric AI agent model employing "work notes" to record and reflect the state throughout task execution. This paper details the model's architecture, featuring worker threads for task oversight, planner modules for task decomposition and planning, and executor modules for performing subtasks using a ReAct-inspired thought-action loop. We provide an exhaustive work state record incorporating plans and outcomes, constituting a comprehensive work journal. Our results show that this model not only improves task execution efficiency but also lays a solid foundation for subsequent task analysis and auditing.
Authors: Adithya Renduchintala, Tugrul Konuk, Oleksii Kuchaiev
We propose Tied-LoRA, a simple paradigm utilizes weight tying and selective training to further increase parameter efficiency of the Low-rank adaptation (LoRA) method. Our investigations include all feasible combinations parameter training/freezing in conjunction with weight tying to identify the optimal balance between performance and the number of trainable parameters. Through experiments covering a variety of tasks and two base language models, we provide analysis revealing trade-offs between efficiency and performance. Our experiments uncovered a particular Tied-LoRA configuration that stands out by demonstrating comparable performance across several tasks while employing only 13~\% percent of parameters utilized by the standard LoRA method.
Authors: Yoonsang Lee, Pranav Atreya, Xi Ye, Eunsol Choi
In-context learning has been applied to knowledge-rich tasks such as question answering. In such scenarios, in-context examples are used to trigger a behaviour in the language model: namely, it should surface information stored in its parametric knowledge. We study the construction of in-context example sets, with a focus on the parametric knowledge of the model regarding in-context examples. We identify 'known' examples, where models can correctly answer from its parametric knowledge, and 'unknown' ones. Our experiments show that prompting with 'unknown' examples decreases the performance, potentially as it encourages hallucination rather than searching its parametric knowledge. Constructing an in-context example set that presents both known and unknown information performs the best across diverse settings. We perform analysis on three multi-answer question answering datasets, which allows us to further study answer set ordering strategies based on the LM's knowledge about each answer. Together, our study sheds lights on how to best construct in-context example sets for knowledge-rich tasks.
Authors: Haofei Yu, Paul Pu Liang, Ruslan Salakhutdinov, Louis-Philippe Morency
Multimodal machine learning, which studies the information and interactions across various input modalities, has made significant advancements in understanding the relationship between images and descriptive text. However, this is just a portion of the potential multimodal interactions seen in the real world and does not include new interactions between conflicting utterances and gestures in predicting sarcasm, for example. Notably, the current methods for capturing shared information often do not extend well to these more nuanced interactions, sometimes performing as low as 50% in binary classification. In this paper, we address this problem via a new approach called MMOE, which stands for a mixture of multimodal interaction experts. Our method automatically classifies data points from unlabeled multimodal datasets by their interaction type and employs specialized models for each specific interaction. Based on our experiments, this approach improves performance on these challenging interactions by more than 10%, leading to an overall increase of 2% for tasks like sarcasm prediction. As a result, interaction quantification provides new insights for dataset analysis and yields simple approaches that obtain state-of-the-art performance.
Authors: Yiqing Xie, Sheng Zhang, Hao Cheng, Zelalem Gero, Cliff Wong, Tristan Naumann, Hoifung Poon
In the evaluation of medical text generation, it is essential to scrutinize each piece of information and ensure the utmost accuracy of the evaluation. Existing evaluation metrics either focus on coarse-level evaluation that assigns one score for the whole generated output or rely on evaluation models trained on general domain, resulting in inaccuracies when adapted to the medical domain. To address these issues, we propose a set of factuality-centric evaluation aspects and design corresponding GPT-4-based metrics for medical text generation. We systematically compare these metrics with existing ones on clinical note generation and medical report summarization tasks, revealing low inter-metric correlation. A comprehensive human evaluation confirms that the proposed GPT-4-based metrics exhibit substantially higher agreement with human judgments than existing evaluation metrics. Our study contributes to the understanding of medical text generation evaluation and offers a more reliable alternative to existing metrics.
Authors: Minbeom Kim, Jahyun Koo, Hwanhee Lee, Joonsuk Park, Hwaran Lee, Kyomin Jung
As large language models become increasingly integrated into daily life, detecting implicit toxicity across diverse contexts is crucial. To this end, we introduce LifeTox, a dataset designed for identifying implicit toxicity within a broad range of advice-seeking scenarios. Unlike existing safety datasets, LifeTox comprises diverse contexts derived from personal experiences through open-ended questions. Experiments demonstrate that RoBERTa fine-tuned on LifeTox matches or surpasses the zero-shot performance of large language models in toxicity classification tasks. These results underscore the efficacy of LifeTox in addressing the complex challenges inherent in implicit toxicity.
Authors: Ramya Ramakrishnan, Ethan Elenberg, Hashan Narangodage, Ryan McDonald
In task-oriented dialogue, a system often needs to follow a sequence of actions, called a workflow, that complies with a set of guidelines in order to complete a task. In this paper, we propose the novel problem of multi-step workflow action prediction, in which the system predicts multiple future workflow actions. Accurate prediction of multiple steps allows for multi-turn automation, which can free up time to focus on more complex tasks. We propose three modeling approaches that are simple to implement yet lead to more action automation: 1) fine-tuning on a training dataset, 2) few-shot in-context learning leveraging retrieval and large language model prompting, and 3) zero-shot graph traversal, which aggregates historical action sequences into a graph for prediction. We show that multi-step action prediction produces features that improve accuracy on downstream dialogue tasks like predicting task success, and can increase automation of steps by 20% without requiring as much feedback from a human overseeing the system.
Authors: Smriti Singh, Cornelia Caragea, Junyi Jessy Li
Situations and events evoke emotions in humans, but to what extent do they inform the prediction of emotion detection models? Prior work in emotion trigger or cause identification focused on training models to recognize events that trigger an emotion. Instead, this work investigates how well human-annotated emotion triggers correlate with features that models deemed salient in their prediction of emotions. First, we introduce a novel dataset EmoTrigger, consisting of 900 social media posts sourced from three different datasets; these were annotated by experts for emotion triggers with high agreement. Using EmoTrigger, we evaluate the ability of large language models (LLMs) to identify emotion triggers, and conduct a comparative analysis of the features considered important for these tasks between LLMs and fine-tuned models. Our analysis reveals that emotion triggers are largely not considered salient features for emotion prediction models, instead there is intricate interplay between various features and the task of emotion detection.
Authors: Ziyi Liu, Isabelle Lee, Yongkang Du, Soumya Sanyal, Jieyu Zhao
Large language models (LLMs) have demonstrated impressive reasoning ability in various language-based tasks. Despite many proposed reasoning methods aimed at enhancing performance in downstream tasks, two fundamental questions persist: Does reasoning genuinely support predictions, and how reliable is the quality of reasoning? In this paper, we propose a framework \textsc{SCORE} to analyze how well LLMs can reason. Specifically, we focus on self-contradictory reasoning, where reasoning does not support the prediction. We find that LLMs often contradict themselves when performing reasoning tasks that involve contextual information and commonsense. The model may miss evidence or use shortcuts, thereby exhibiting self-contradictory behaviors. We also employ the Point-of-View (POV) method, which probes models to generate reasoning from multiple perspectives, as a diagnostic tool for further analysis. We find that though LLMs may appear to perform well in one-perspective settings, they fail to stabilize such behavior in multi-perspectives settings. Even for correct predictions, the reasoning may be messy and incomplete, and LLMs can easily be led astray from good reasoning. \textsc{SCORE}'s results underscore the lack of robustness required for trustworthy reasoning and the urgency for further research to establish best practices for a comprehensive evaluation of reasoning beyond accuracy-based metrics.
Authors: Yanai Elazar, Bhargavi Paranjape, Hao Peng, Sarah Wiegreffe, Khyathi Raghavi, Vivek Srikumar, Sameer Singh, Noah A. Smith
The inevitable appearance of spurious correlations in training datasets hurts the generalization of NLP models on unseen data. Previous work has found that datasets with paired inputs are prone to correlations between a specific part of the input (e.g., the hypothesis in NLI) and the label; consequently, models trained only on those outperform chance. Are these correlations picked up by models trained on the full input data? To address this question, we propose a new evaluation method, Counterfactual Attentiveness Test (CAT). CAT uses counterfactuals by replacing part of the input with its counterpart from a different example (subject to some restrictions), expecting an attentive model to change its prediction. Using CAT, we systematically investigate established supervised and in-context learning models on ten datasets spanning four tasks: natural language inference, reading comprehension, paraphrase detection, and visual & language reasoning. CAT reveals that reliance on such correlations is mainly data-dependent. Surprisingly, we find that GPT3 becomes less attentive with an increased number of demonstrations, while its accuracy on the test data improves. Our results demonstrate that augmenting training or demonstration data with counterfactuals is effective in improving models' attentiveness. We show that models' attentiveness measured by CAT reveals different conclusions from solely measuring correlations in data.
Authors: Shivanshu Gupta, Clemens Rosenbaum, Ethan R. Elenberg
Large language models (LLMs) have the ability to perform in-context learning (ICL) of new tasks by conditioning on prompts comprising a few task examples. This work studies the problem of selecting the best examples given a candidate pool to improve ICL performance on given a test input. Existing approaches either require training with feedback from a much larger LLM or are computationally expensive. We propose a novel metric, GistScore, based on Example Gisting, a novel approach for training example retrievers for ICL using an attention bottleneck via Gisting, a recent technique for compressing task instructions. To tradeoff performance with ease of use, we experiment with both fine-tuning gist models on each dataset and multi-task training a single model on a large collection of datasets. On 21 diverse datasets spanning 9 tasks, we show that our fine-tuned models get state-of-the-art ICL performance with 20% absolute average gain over off-the-shelf retrievers and 7% over the best prior methods. Our multi-task model generalizes well out-of-the-box to new task categories, datasets, and prompt templates with retrieval speeds that are consistently thousands of times faster than the best prior training-free method.
Authors: Wang Zhu, Alekh Agarwal, Mandar Joshi, Robin Jia, Jesse Thomason, Kristina Toutanova
Understanding visually situated language requires recognizing text and visual elements, and interpreting complex layouts. State-of-the-art methods commonly use specialized pre-processing tools, such as optical character recognition (OCR) systems, that map document image inputs to extracted information in the space of textual tokens, and sometimes also employ large language models (LLMs) to reason in text token space. However, the gains from external tools and LLMs come at the cost of increased computational and engineering complexity. In this paper, we ask whether small pretrained image-to-text models can learn selective text or layout recognition and reasoning as an intermediate inference step in an end-to-end model for pixel-level visual language understanding. We incorporate the outputs of such OCR tools, LLMs, and larger multimodal models as intermediate ``rationales'' on training data, and train a small student model to predict both rationales and answers for input questions based on those training examples. A student model based on Pix2Struct (282M parameters) achieves consistent improvements on three visual document understanding benchmarks representing infographics, scanned documents, and figures, with improvements of more than 4\% absolute over a comparable Pix2Struct model that predicts answers directly.
Authors: Yuling Gu, Oyvind Tafjord, Peter Clark
While LLMs can provide reasoned explanations along with their answers, the nature and quality of those explanations are still poorly understood. In response, our goal is to define a detailed way of characterizing the explanation capabilities of modern models and to create a nuanced, interpretable explanation evaluation tool that can generate such characterizations automatically, without relying on expensive API calls or human annotations. Our approach is to (a) define the new task of explanation critiquing - identifying and categorizing any main flaw in an explanation and providing suggestions to address the flaw, (b) create a sizeable, human-verified dataset for this task, and (c) train an open-source, automatic critiquing model (called Digital Socrates) using this data. Through quantitative and qualitative analysis, we demonstrate how Digital Socrates is useful for revealing insights about student models by examining their reasoning chains, and how it can provide high-quality, nuanced, automatic evaluation of those model explanations for the first time. Digital Socrates thus fills an important gap in evaluation tools for understanding and improving the explanation behavior of models.
Authors: Ting-Rui Chiang, Xinyan Velocity Yu, Joshua Robinson, Ollie Liu, Isabelle Lee, Dani Yogatama
Augmenting a language model (LM) with $k$-nearest neighbors (kNN) retrieval on its training data alone can decrease its perplexity, though the underlying reasons for this remains elusive. In this work, we first rule out one previously posited possibility -- the "softmax bottleneck." We further identify the MLP hurdle phenomenon, where the final MLP layer in LMs may impede LM optimization early on. We explore memorization and generalization in language models with two new datasets, where advanced model like GPT-3.5-turbo find generalizing to irrelevant information in the training data challenging. However, incorporating kNN retrieval to vanilla GPT-2 117M can consistently improve performance in this setting.
Authors: Yun-Shiuan Chuang, Agam Goyal, Nikunj Harlalka, Siddharth Suresh, Robert Hawkins, Sijia Yang, Dhavan Shah, Junjie Hu, Timothy T. Rogers
Accurately simulating human opinion dynamics is crucial for understanding a variety of societal phenomena, including polarization and the spread of misinformation. However, the agent-based models (ABMs) commonly used for such simulations lack fidelity to human behavior. We propose a new approach to simulating opinion dynamics based on populations of Large Language Models (LLMs). Our findings reveal a strong inherent bias in LLM agents towards accurate information, leading to consensus in line with scientific reality. However, this bias limits the simulation of individuals with resistant views on issues like climate change. After inducing confirmation bias through prompt engineering, we observed opinion fragmentation in line with existing agent-based research. These insights highlight the promise and limitations of LLM agents in this domain and suggest a path forward: refining LLMs with real-world discourse to better simulate the evolution of human beliefs.
Authors: Kazuma Hashimoto, Karthik Raman, Michael Bendersky
In-Context Learning (ICL) is an emergent capability of Large Language Models (LLMs). Only a few demonstrations enable LLMs to be used as blackbox for new tasks. Previous studies have shown that using LLMs' outputs as labels is effective in training models to select demonstrations. Such a label is expected to estimate utility of a demonstration in ICL; however, it has not been well understood how different labeling strategies affect results on target tasks. This paper presents an analysis on different utility functions by focusing on LLMs' output probability given ground-truth output, and task-specific reward given LLMs' prediction. Unlike the previous work, we introduce a novel labeling method, incremental utility, which estimates how much incremental knowledge is brought into the LLMs by a demonstration. We conduct experiments with instruction-tuned LLMs on binary/multi-class classification, segmentation, and translation across Arabic, English, Finnish, Japanese, and Spanish. Our results show that (1) the probability is effective when the probability values are distributed across the whole value range (on the classification tasks), and (2) the downstream metric is more robust when nuanced reward values are provided with long outputs (on the segmentation and translation tasks). We then show that the proposed incremental utility further helps ICL by contrasting how the LLMs perform with and without the demonstrations.
Authors: Nakyeong Yang, Taegwan Kang, Kyomin Jung
Large language models (LLMs) executing tasks through instruction-based prompts often face challenges stemming from distribution differences between user instructions and training instructions. This leads to distractions and biases, especially when dealing with inconsistent dynamic labels. In this paper, we introduces a novel bias mitigation method, CRISPR, designed to alleviate instruction-label biases in LLMs. CRISPR utilizes attribution methods to identify bias neurons influencing biased outputs and employs pruning to eliminate the bias neurons. Experimental results demonstrate the method's effectiveness in mitigating biases in instruction-based prompting, enhancing language model performance on social bias benchmarks without compromising pre-existing knowledge. CRISPR proves highly practical, model-agnostic, offering flexibility in adapting to evolving social biases.
Authors: Yanchen Liu, Mingyu Derek Ma, Wenna Qin, Azure Zhou, Jiaao Chen, Weiyan Shi, Wei Wang, Diyi Yang
Susceptibility to misinformation describes the extent to believe unverifiable claims, which is hidden in people's mental process and infeasible to observe. Existing susceptibility studies heavily rely on the self-reported beliefs, making any downstream applications on susceptability hard to scale. To address these limitations, in this work, we propose a computational model to infer users' susceptibility levels given their activities. Since user's susceptibility is a key indicator for their reposting behavior, we utilize the supervision from the observable sharing behavior to infer the underlying susceptibility tendency. The evaluation shows that our model yields estimations that are highly aligned with human judgment on users' susceptibility level comparisons. Building upon such large-scale susceptibility labeling, we further conduct a comprehensive analysis of how different social factors relate to susceptibility. We find that political leanings and psychological factors are associated with susceptibility in varying degrees.
Authors: Yuhao Wu, Tongjun Shi, Karthick Sharma, Chun Wei Seah, Shuhao Zhang
Large Language Models (LLMs) serve as repositories of extensive world knowledge, enabling them to perform tasks such as question-answering and fact-checking. However, this knowledge can become obsolete as global contexts change. In this paper, we introduce a novel problem in the realm of continual learning: Online Continual Knowledge Learning (OCKL). This problem formulation aims to manage the dynamic nature of world knowledge in LMs under real-time constraints. We propose a new benchmark and evaluation metric designed to measure both the rate of new knowledge acquisition and the retention of previously learned knowledge. Our empirical evaluation, conducted using a variety of state-of-the-art methods, establishes robust base-lines for OCKL. Our results reveal that existing continual learning approaches are unfortunately insufficient for tackling the unique challenges posed by OCKL. We identify key factors that influence the trade-off between knowledge acquisition and retention, thereby advancing our understanding of how to train LMs in a continually evolving environment.
Authors: Arkil Patel, Siva Reddy, Dzmitry Bahdanau, Pradeep Dasigi
Contemporary Large Language Models (LLMs) exhibit a high degree of code generation and comprehension capability. A particularly promising area is their ability to interpret code modules from unfamiliar libraries for solving user-instructed tasks. Recent work has shown that large proprietary LLMs can learn novel library usage in-context from demonstrations. These results raise several open questions: whether demonstrations of library usage is required, whether smaller (and more open) models also possess such capabilities, etc. In this work, we take a broader approach by systematically evaluating a diverse array of LLMs across three scenarios reflecting varying levels of domain specialization to understand their abilities and limitations in generating code based on libraries defined in-context. Our results show that even smaller open-source LLMs like Llama-2 and StarCoder demonstrate an adept understanding of novel code libraries based on specification presented in-context. Our findings further reveal that LLMs exhibit a surprisingly high proficiency in learning novel library modules even when provided with just natural language descriptions or raw code implementations of the functions, which are often cheaper to obtain than demonstrations. Overall, our results pave the way for harnessing LLMs in more adaptable and dynamic coding environments.
Authors: Jiongxiao Wang, Junlin Wu, Muhao Chen, Yevgeniy Vorobeychik, Chaowei Xiao
Reinforcement Learning with Human Feedback (RLHF) is a methodology designed to align Large Language Models (LLMs) with human preferences, playing an important role in LLMs alignment. Despite its advantages, RLHF relies on human annotators to rank the text, which can introduce potential security vulnerabilities if any adversarial annotator (i.e., attackers) manipulates the ranking score by up-ranking any malicious text to steer the LLM adversarially. To assess the red-teaming of RLHF against human preference data poisoning, we propose RankPoison, a poisoning attack method on candidates' selection of preference rank flipping to reach certain malicious behaviors (e.g., generating longer sequences, which can increase the computational cost). With poisoned dataset generated by RankPoison, we can perform poisoning attacks on LLMs to generate longer tokens without hurting the original safety alignment performance. Moreover, applying RankPoison, we also successfully implement a backdoor attack where LLMs can generate longer answers under questions with the trigger word. Our findings highlight critical security challenges in RLHF, underscoring the necessity for more robust alignment methods for LLMs.
Authors: Yidan Sun, Qin Chao, Boyang Li
Psychological research suggests the central role of event causality in human story understanding. Further, event causality has been heavily utilized in symbolic story generation. However, few machine learning systems for story understanding employ event causality, partially due to the lack of reliable methods for identifying open-world causal event relations. Leveraging recent progress in large language models (LLMs), we present the first method for event causality identification that leads to material improvements in computational story understanding. We design specific prompts for extracting event causal relations from GPT. Against human-annotated event causal relations in the GLUCOSE dataset, our technique performs on par with supervised models, while being easily generalizable to stories of different types and lengths. The extracted causal relations lead to 5.7\% improvements on story quality evaluation and 8.7\% on story video-text alignment. Our findings indicate enormous untapped potential for event causality in computational story understanding.
Authors: Yaxin Zhu, Hamed Zamani
This paper focuses on the task of Extreme Multi-Label Classification (XMC) whose goal is to predict multiple labels for each instance from an extremely large label space. While existing research has primarily focused on fully supervised XMC, real-world scenarios often lack complete supervision signals, highlighting the importance of zero-shot settings. Given the large label space, utilizing in-context learning approaches is not trivial. We address this issue by introducing In-Context Extreme Multilabel Learning (ICXML), a two-stage framework that cuts down the search space by generating a set of candidate labels through incontext learning and then reranks them. Extensive experiments suggest that ICXML advances the state of the art on two diverse public benchmarks.
Authors: Siru Ouyang, Zhuosheng Zhang, Bing Yan, Xuan Liu, Jiawei Han, Lianhui Qin
This paper studies the problem of solving complex chemistry problems with large language models (LLMs). Despite the extensive general knowledge in LLMs (such as GPT-4), they struggle with chemistry reasoning that requires faithful grounded reasoning with diverse chemical knowledge and an integrative understanding of chemical interactions. We propose InstructChem, a new structured reasoning approach that substantially boosts the LLMs' chemical reasoning capabilities. InstructChem explicitly decomposes the reasoning into three critical phrases, including chemical formulae generation by LLMs that offers the basis for subsequent grounded reasoning, step-by-step reasoning that makes multi-step derivations with the identified formulae for a preliminary answer, and iterative review-and-refinement that steers LLMs to progressively revise the previous phases for increasing confidence, leading to the final high-confidence answer. We conduct extensive experiments on four different chemistry challenges, including quantum chemistry, quantum mechanics, physical chemistry, and chemistry kinetics. Our approach significantly enhances GPT-4 on chemistry reasoning, yielding an 8% average absolute improvement and a 30% peak improvement. We further use the generated reasoning by GPT-4 to fine-tune smaller LMs (e.g., Vicuna) and observe strong improvement of the smaller LMs. This validates our approach and enables LLMs to generate high-quality reasoning.
Authors: Yun-Shiuan Chuang, Yi Wu, Dhruv Gupta, Rheeya Uppaal, Ananya Kumar, Luhang Sun, Makesh Narsimhan Sreedhar, Sijia Yang, Timothy T. Rogers, Junjie Hu
Adapting pre-trained language models (PLMs) for time-series text classification amidst evolving domain shifts (EDS) is critical for maintaining accuracy in applications like stance detection. This study benchmarks the effectiveness of evolving domain adaptation (EDA) strategies, notably self-training, domain-adversarial training, and domain-adaptive pretraining, with a focus on an incremental self-training method. Our analysis across various datasets reveals that this incremental method excels at adapting PLMs to EDS, outperforming traditional domain adaptation techniques. These findings highlight the importance of continually updating PLMs to ensure their effectiveness in real-world applications, paving the way for future research into PLM robustness against the natural temporal evolution of language.
Authors: Yun-Shiuan Chuang, Siddharth Suresh, Nikunj Harlalka, Agam Goyal, Robert Hawkins, Sijia Yang, Dhavan Shah, Junjie Hu, Timothy T. Rogers
This study investigates the potential of Large Language Models (LLMs) to simulate human group dynamics, particularly within politically charged contexts. We replicate the Wisdom of Partisan Crowds phenomenon using LLMs to role-play as Democrat and Republican personas, engaging in a structured interaction akin to human group study. Our approach evaluates how agents' responses evolve through social influence. Our key findings indicate that LLM agents role-playing detailed personas and without Chain-of-Thought (CoT) reasoning closely align with human behaviors, while having CoT reasoning hurts the alignment. However, incorporating explicit biases into agent prompts does not necessarily enhance the wisdom of partisan crowds. Moreover, fine-tuning LLMs with human data shows promise in achieving human-like behavior but poses a risk of overfitting certain behaviors. These findings show the potential and limitations of using LLM agents in modeling human group phenomena.
Authors: Yuhang Li, Yihan Wang, Zhouxing Shi, Cho-Jui Hsieh
The strong general capabilities of Large Language Models (LLMs) bring potential ethical risks if they are unrestrictedly accessible to malicious users. Token-level watermarking inserts watermarks in the generated texts by altering the token probability distributions with a private random number generator seeded by its prefix tokens. However, this watermarking algorithm alters the logits during generation, which can lead to a downgraded text quality if it chooses to promote tokens that are less relevant given the input. In this work, we propose to improve the quality of texts generated by a watermarked language model by Watermarking with Importance Scoring (WIS). At each generation step, we estimate the importance of the token to generate, and prevent it from being impacted by watermarking if it is important for the semantic correctness of the output. We further propose three methods to predict importance scoring, including a perturbation-based method and two model-based methods. Empirical experiments show that our method can generate texts with better quality with comparable level of detection rate.
Authors: Maria Antoniak, Joel Mire, Maarten Sap, Elliott Ash, Andrew Piper
People share stories online for a myriad of purposes, whether as a means of self-disclosure, processing difficult personal experiences, providing needed information or entertainment, or persuading others to share their beliefs. Better understanding of online storytelling can illuminate the dynamics of social movements, sensemaking practices, persuasion strategies, and more. However, unlike other media such as books and visual content where the narrative nature of the content is often overtly signaled at the document level, studying storytelling in online communities is challenging due to the mixture of storytelling and non-storytelling behavior, which can be interspersed within documents and across diverse topics and settings. We introduce a codebook and create the Storytelling in Online Communities Corpus, an expert-annotated dataset of 502 English-language posts and comments with labeled story and event spans. Using our corpus, we train and evaluate an online story detection model, which we use to investigate the role storytelling of in different social contexts. We identify distinctive features of online storytelling, the prevalence of storytelling among different communities, and the conversational patterns of storytelling.
Authors: Hanning Zhang, Shizhe Diao, Yong Lin, Yi R. Fung, Qing Lian, Xingyao Wang, Yangyi Chen, Heng Ji, Tong Zhang
Large language models (LLMs) have revolutionized numerous domains with their impressive performance but still face their challenges. A predominant issue is the propensity for these models to generate non-existent facts, a concern termed hallucination. Our research is motivated by the observation that previous instruction tuning methods force the model to complete a sentence no matter whether the model knows the knowledge or not. When the question is out of the parametric knowledge, it will try to make up something and fail to indicate when it lacks knowledge. In this paper, we present a new approach called Refusal-Aware Instruction Tuning (R-Tuning). This approach is formalized by first identifying the knowledge gap between parametric knowledge and the instruction tuning data. Then, we construct the refusal-aware data based on the knowledge intersection, to tune LLMs to refrain from responding to questions beyond its parametric knowledge. Experimental results demonstrate this new instruction tuning approach effectively improves a model's ability to answer known questions and refrain from answering unknown questions. Furthermore, when tested on out-of-domain datasets, the refusal ability was found to be a meta-skill that could be generalized to other tasks. Further analysis surprisingly finds that learning the uncertainty during training displays a better ability to estimate uncertainty than uncertainty-based testing. Our code will be released at https://github.com/shizhediao/R-Tuning.
Authors: Yufei Tian, Abhilasha Ravichander, Lianhui Qin, Ronan Le Bras, Raja Marjieh, Nanyun Peng, Yejin Choi, Thomas L. Griffiths, Faeze Brahman
We explore the creative problem-solving capabilities of modern large language models (LLMs) in a constrained setting. The setting requires circumventing a cognitive bias known in psychology as ''functional fixedness'' to use familiar objects in innovative or unconventional ways. To this end, we create MacGyver, an automatically generated dataset consisting of 1,600 real-world problems that deliberately trigger functional fixedness and require thinking 'out-of-the-box'. We then present our collection of problems to both LLMs and humans to compare and contrast their problem-solving abilities. We show that MacGyver is challenging for both groups, but in unique and complementary ways. For example, humans typically excel in solving problems that they are familiar with but may struggle with tasks requiring domain-specific knowledge, leading to a higher variance. On the other hand, LLMs, being exposed to a variety of highly specialized knowledge, attempt broader problems but are prone to overconfidence and propose actions that are physically infeasible or inefficient. We also provide a detailed error analysis of LLMs, and demonstrate the potential of enhancing their problem-solving ability with novel prompting techniques such as iterative step-wise reflection and divergent-convergent thinking. This work provides insight into the creative problem-solving capabilities of humans and AI and illustrates how psychological paradigms can be extended into large-scale tasks for comparing humans and machines.
Authors: Zonghai Yao, Ahmed Jaafar, Beining Wang, Yue Zhu, Zhichao Yang, Hong Yu
This study examines the effect of prompt engineering on the performance of Large Language Models (LLMs) in clinical note generation. We introduce an Automatic Prompt Optimization (APO) framework to refine initial prompts and compare the outputs of medical experts, non-medical experts, and APO-enhanced GPT3.5 and GPT4. Results highlight GPT4 APO's superior performance in standardizing prompt quality across clinical note sections. A human-in-the-loop approach shows that experts maintain content quality post-APO, with a preference for their own modifications, suggesting the value of expert customization. We recommend a two-phase optimization process, leveraging APO-GPT4 for consistency and expert input for personalization.
Authors: Zihao He, Siyi Guo, Ashwin Rao, Kristina Lerman
Social media platforms are rife with politically charged discussions. Therefore, accurately deciphering and predicting partisan biases using Large Language Models (LLMs) is increasingly critical. In this study, we address the challenge of understanding political bias in digitized discourse using LLMs. While traditional approaches often rely on finetuning separate models for each political faction, our work innovates by employing a singular, instruction-tuned LLM to reflect a spectrum of political ideologies. We present a comprehensive analytical framework, consisting of Partisan Bias Divergence Assessment and Partisan Class Tendency Prediction, to evaluate the model's alignment with real-world political ideologies in terms of stances, emotions, and moral foundations. Our findings reveal the model's effectiveness in capturing emotional and moral nuances, albeit with some challenges in stance detection, highlighting the intricacies and potential for refinement in NLP tools for politically sensitive contexts. This research contributes significantly to the field by demonstrating the feasibility and importance of nuanced political understanding in LLMs, particularly for applications requiring acute awareness of political bias.
Authors: Andrew Blair-Stanek, Nils Holzenberger, Benjamin Van Durme
We find that the best publicly available LLMs like GPT-4 and PaLM 2 currently perform poorly at basic text handling required of lawyers or paralegals, such as looking up the text at a line of a witness deposition or at a subsection of a contract. We introduce a benchmark to quantify this poor performance, which casts into doubt LLMs' current reliability as-is for legal practice. Finetuning for these tasks brings an older LLM to near-perfect performance on our test set and also raises performance on a related legal task. This stark result highlights the need for more domain expertise in LLM training.
Authors: Ashim Gupta, Rishanth Rajendhran, Nathan Stringham, Vivek Srikumar, Ana Marasović
Are the longstanding robustness issues in NLP resolved by today's larger and more performant models? To address this question, we conduct a thorough investigation using 19 models of different sizes spanning different architectural choices and pretraining objectives. We conduct evaluations using (a) OOD and challenge test sets, (b) CheckLists, (c) contrast sets, and (d) adversarial inputs. Our analysis reveals that not all OOD tests provide further insight into robustness. Evaluating with CheckLists and contrast sets shows significant gaps in model performance; merely scaling models does not make them sufficiently robust. Finally, we point out that current approaches for adversarial evaluations of models are themselves problematic: they can be easily thwarted, and in their current forms, do not represent a sufficiently deep probe of model robustness. We conclude that not only is the question of robustness in NLP as yet unresolved, but even some of the approaches to measure robustness need to be reassessed.
Authors: Wei-Rui Chen, Ife Adebara, Khai Duy Doan, Qisheng Liao, Muhammad Abdul-Mageed
Recently, ChatGPT has emerged as a powerful NLP tool that can carry out several tasks. However, the range of languages ChatGPT can handle remains largely a mystery. In this work, we investigate ChatGPT's language identification abilities. For this purpose, we compile Babel-670, a benchmark comprising $670$ languages representing $23$ language families. Languages in Babel-670 run the gamut between the very high-resource to the very low-resource and are spoken in five continents. We then study ChatGPT's (both GPT-3.5 and GPT-4) ability to (i) identify both language names and language codes (ii) under both zero- and few-shot conditions (iii) with and without provision of label set. When compared to smaller finetuned language identification tools, we find that ChatGPT lags behind. Our empirical analysis shows the reality that ChatGPT still resides in a state of potential enhancement before it can sufficiently serve diverse communities.
Authors: Bangzheng Li, Ben Zhou, Fei Wang, Xingyu Fu, Dan Roth, Muhao Chen
Despite the recent advancement in large language models (LLMs) and their high performances across numerous benchmarks, recent research has unveiled that LLMs suffer from hallucinations and unfaithful reasoning. This work studies a specific type of hallucination induced by semantic associations. Specifically, we investigate to what extent LLMs take shortcuts from certain keyword/entity biases in the prompt instead of following the correct reasoning path. To quantify this phenomenon, we propose a novel probing method and benchmark called EureQA. We start from questions that LLMs will answer correctly with utmost certainty, and mask the important entity with evidence sentence recursively, asking models to find masked entities according to a chain of evidence before answering the question.
During the construction of the evidence, we purposefully replace semantic clues (entities) that may lead to the correct answer with distractor clues (evidence) that will not directly lead to the correct answer but require a chain-like reasoning process. We evaluate if models can follow the correct reasoning chain instead of short-cutting through distractor clues. We find that existing LLMs lack the necessary capabilities to follow correct reasoning paths and resist the attempt of greedy shortcuts. We show that the distractor semantic associations often lead to model hallucination, which is strong evidence that questions the validity of current LLM reasoning.
Authors: Andor Diera, Abdelhalim Dahou, Lukas Galke, Fabian Karl, Florian Sihler, Ansgar Scherp
Language models can serve as a valuable tool for software developers to increase productivity. Large generative models can be used for code generation and code completion, while smaller encoder-only models are capable of performing code search tasks using natural language queries.These capabilities are heavily influenced by the quality and diversity of the available training data. Source code datasets used for training usually focus on the most popular languages and testing is mostly conducted on the same distributions, often overlooking low-resource programming languages. Motivated by the NLP generalization taxonomy proposed by Hupkes et.\,al., we propose a new benchmark dataset called GenCodeSearchNet (GeCS) which builds upon existing natural language code search datasets to systemically evaluate the programming language understanding generalization capabilities of language models. As part of the full dataset, we introduce a new, manually curated subset StatCodeSearch that focuses on R, a popular but so far underrepresented programming language that is often used by researchers outside the field of computer science. For evaluation and comparison, we collect several baseline results using fine-tuned BERT-style models and GPT-style large language models in a zero-shot setting.
Authors: Thi-Nhung Nguyen, Hoang Ngo, Kiem-Hieu Nguyen, Tuan-Dung Cao
Our work addresses the problem of unsupervised Aspect Category Detection using a small set of seed words. Recent works have focused on learning embedding spaces for seed words and sentences to establish similarities between sentences and aspects. However, aspect representations are limited by the quality of initial seed words, and model performances are compromised by noise. To mitigate this limitation, we propose a simple framework that automatically enhances the quality of initial seed words and selects high-quality sentences for training instead of using the entire dataset. Our main concepts are to add a number of seed words to the initial set and to treat the task of noise resolution as a task of augmenting data for a low-resource task. In addition, we jointly train Aspect Category Detection with Aspect Term Extraction and Aspect Term Polarity to further enhance performance. This approach facilitates shared representation learning, allowing Aspect Category Detection to benefit from the additional guidance offered by other tasks. Extensive experiments demonstrate that our framework surpasses strong baselines on standard datasets.
Authors: Nikolay Bogoychev, Pinzhen Chen, Barry Haddow, Alexandra Birch
Large language model (LLM) inference is computation and memory intensive, so we adapt lexical shortlisting to it hoping to improve both. While lexical shortlisting is well-explored in tasks like machine translation, it requires modifications before being suitable for LLMs as the intended applications vary significantly. Our work studies two heuristics to shortlist sub-vocabulary at LLM inference time: Unicode-based script filtering and corpus-based selection. We explore different LLM families and sizes, and we find that lexical shortlisting can reduce the memory usage of some models by nearly 50\% and has an upper bound of 25\% improvement in generation speed. In this pilot study, we also identify the drawbacks of such vocabulary selection methods and propose avenues for future research.
Authors: Athul Paul Jacob, Gabriele Farina, Jacob Andreas
We present a model of pragmatic language understanding, where utterances are produced and understood by searching for regularized equilibria of signaling games. In this model (which we call ReCo, for Regularized Conventions), speakers and listeners search for contextually appropriate utterance--meaning mappings that are both close to game-theoretically optimal conventions and close to a shared, ''default'' semantics. By characterizing pragmatic communication as equilibrium search, we obtain principled sampling algorithms and formal guarantees about the trade-off between communicative success and naturalness. Across several datasets capturing real and idealized human judgments about pragmatic implicatures, ReCo matches or improves upon predictions made by best response and rational speech act models of language understanding.
Authors: Bangzhao Shu, Lechen Zhang, Minje Choi, Lavinia Dunagan, Dallas Card, David Jurgens
The versatility of Large Language Models (LLMs) on natural language understanding tasks has made them popular for research in social sciences. In particular, to properly understand the properties and innate personas of LLMs, researchers have performed studies that involve using prompts in the form of questions that ask LLMs of particular opinions. In this study, we take a cautionary step back and examine whether the current format of prompting enables LLMs to provide responses in a consistent and robust manner. We first construct a dataset that contains 693 questions encompassing 39 different instruments of persona measurement on 115 persona axes. Additionally, we design a set of prompts containing minor variations and examine LLM's capabilities to generate accurate answers, as well as consistency variations to examine their consistency towards simple perturbations such as switching the option order. Our experiments on 15 different open-source LLMs reveal that even simple perturbations are sufficient to significantly downgrade a model's question-answering ability, and that most LLMs have low negation consistency. Our results suggest that the currently widespread practice of prompting is insufficient to accurately capture model perceptions, and we discuss potential alternatives to improve such issues.
Authors: Linyong Nan, Ellen Zhang, Weijin Zou, Yilun Zhao, Wenfei Zhou, Arman Cohan
This study introduces a new long-form database question answering dataset designed to evaluate how Large Language Models (LLMs) interact with a SQL interpreter. The task necessitates LLMs to strategically generate multiple SQL queries to retrieve sufficient data from a database, to reason with the acquired context, and to synthesize them into a comprehensive analytical narrative. Our findings highlight that this task poses great challenges even for the state-of-the-art GPT-4 model. We propose and evaluate two interaction strategies, and provide a fine-grained analysis of the individual stages within the interaction. A key discovery is the identification of two primary bottlenecks hindering effective interaction: the capacity for planning and the ability to generate multiple SQL queries. To address the challenge of accurately assessing answer quality, we introduce a multi-agent evaluation framework that simulates the academic peer-review process, enhancing the precision and reliability of our evaluations. This framework allows for a more nuanced understanding of the strengths and limitations of current LLMs in complex retrieval and reasoning tasks.
Authors: Fei Yu, Anningzhe Gao, Benyou Wang
Large language models (LLMs) often struggle with maintaining accuracy across a sequence of intermediate reasoning steps in mathematical reasoning, leading to error propagation that undermines the final result. The current methodology to mitigate this issue primarily involves using a verifier model to assess the correctness of generated solution candidates, focusing either on the overall reasoning path or on an incomplete reasoning path. By rethinking this approach, we argue that assessing potentials of incomplete reasoning paths could be more advantageous as it guides towards correct final answers, transforming the task into a \textit{planning} problem. Our proposed verifier, the Outcome-supervision Value Model (OVM), employs outcome supervision for training, offering an efficient and intuitive method for \textit{planning} by prioritizing steps that lead to accurate conclusions over mere per-step correctness. Furthermore, the OVM eschews the need for labor-intensive annotations on step-level correctness, enhancing its scalability. Our experiments on two multi-step mathematical reasoning datasets, GSM8K and Game of 24, demonstrate the superior performance of the OVM model. Notably, in GSM8K, our \textbf{OVM-7B model achieves state-of-the-art results among LLMs up to 13B parameters}; especially it does not utilize GPT-4 or code execution. These findings offer a novel perspective on the role of outcome supervision in training verifiers for multi-step reasoning tasks and provide theoretical justification for its advantage in value estimation for planning.
Authors: Huaman Sun, Jiaxin Pei, Minje Choi, David Jurgens
Human perception of language depends on personal backgrounds like gender and ethnicity. While existing studies have shown that large language models (LLMs) hold values that are closer to certain societal groups, it is unclear whether their prediction behaviors on subjective NLP tasks also exhibit a similar bias. In this study, leveraging the POPQUORN dataset which contains annotations of diverse demographic backgrounds, we conduct a series of experiments on four popular LLMs to investigate their capability to understand group differences and potential biases in their predictions for politeness and offensiveness. We find that for both tasks, model predictions are closer to the labels from White and female participants. We further explore prompting with the target demographic labels and show that including the target demographic in the prompt actually worsens the model's performance. More specifically, when being prompted to respond from the perspective of "Black" and "Asian" individuals, models show lower performance in predicting both overall scores as well as the scores from corresponding groups. Our results suggest that LLMs hold gender and racial biases for subjective NLP tasks and that demographic-infused prompts alone may be insufficient to mitigate such effects. Code and data are available at https://github.com/Jiaxin-Pei/LLM-Group-Bias.
Authors: Genglin Liu, Xingyao Wang, Lifan Yuan, Yangyi Chen, Hao Peng
Large Language Models (LLMs) often struggle when faced with situations where they lack the prerequisite knowledge to generate a sensical response. In these cases, models tend to fabricate and hallucinate, rather than appropriately signaling uncertainty as humans would. This behavior misaligns with human conversational norms and presents challenges surrounding responsible and ethical AI development. This work aims to systematically investigate LLMs' behaviors in such situations. We curate an adversarial question-answering benchmark containing unanswerable questions targeting information absent from the LLM's training data. Concretely, these unanswerable questions contain non-existent concepts or false premises. When presented with such unanswerable questions, an LLM should appropriately convey uncertainty, and be able to challenge the premise and refuse to generate a response. While facing answerable valid questions, a model should demonstrate a positive correlation between accuracy and confidence. Using a model-agnostic unified confidence elicitation approach, we observe that LLMs that have gone through instruction finetuning and reinforcement learning from human feedback (RLHF) perform significantly better than their counterparts that do not. Moreover, uncertainty expression 1 through our elicitation method does not always stay consistent with the perceived confidence of the direct response of an LLM. Our findings call for further research into teaching LLMs to proactively and reliably express uncertainty.
Authors: Yipei Xu, Dakuan Lu, Jiaqing Liang, Xintao Wang, Yipeng Geng, Yingsi Xin, Hengkui Wu, Ken Chen, ruiji zhang, Yanghua Xiao
Pre-trained language models (PLMs) have established the new paradigm in the field of NLP. For more powerful PLMs, one of the most popular and successful way is to continuously scale up sizes of the models and the pre-training corpora. These large corpora are generally obtained by converging smaller ones from multiple sources, they are thus growing increasingly diverse. However, the side-effects of these colossal converged corpora remain understudied. In this paper, we identify the disadvantage of heterogeneous corpora from multiple sources for pre-training PLMs. Towards coordinated pre-training on diverse corpora, we further propose source prompts (SP), which explicitly prompt the model of the data source at the pre-training and fine-tuning stages. Results of extensive experiments demonstrate that PLMs pre-trained with SP on diverse corpora gain significant improvement in various downstream tasks.
Authors: Xinliang Frederick Zhang, Winston Wu, Nick Beauchamp, Lu Wang
News media employ moral language to create memorable stories, and readers often engage with the content that align with their values. Moral theories have been applied to news analysis studying moral values in isolation, while the intricate dynamics among participating entities in shaping moral events have been overlooked. This is mainly due to the use of obscure language to conceal evident ideology and values, coupled with the insufficient moral reasoning capability in most existing NLP systems, where LLMs are no exception. To study this phenomenon, we first annotate a new dataset, MORAL EVENTS, consisting of 5,494 structured annotations on 474 news articles by diverse US media across the political spectrum. We further propose MOKA, a moral event extraction framework with MOral Knowledge Augmentation, that leverages knowledge derived from moral words and moral scenarios. Experimental results show that MOKA outperforms competitive baselines across three moral event understanding tasks. Further analyses illuminate the selective reporting of moral events by media outlets of different ideological leanings, suggesting the significance of event-level morality analysis in news. Our datasets and codebase are available at https://github.com/launchnlp/MOKA.
Authors: Alexander Spangher, Emilio Ferrara, Ben Welsh, Nanyun Peng, Serdar Tumgoren, Jonathan May
Journalists must find stories in huge amounts of textual data (e.g. leaks, bills, press releases) as part of their jobs: determining when and why text becomes news can help us understand coverage patterns and help us build assistive tools. Yet, this is challenging because very few labelled links exist, language use between corpora is very different, and text may be covered for a variety of reasons. In this work we focus on news coverage of local public policy in the San Francisco Bay Area by the San Francisco Chronicle. First, we gather news articles, public policy documents and meeting recordings and link them using probabilistic relational modeling, which we show is a low-annotation linking methodology that outperforms other retrieval-based baselines. Second, we define a new task: newsworthiness prediction, to predict if a policy item will get covered. We show that different aspects of public policy discussion yield different newsworthiness signals. Finally we perform human evaluation with expert journalists and show our systems identify policies they consider newsworthy with 68% F1 and our coverage recommendations are helpful with an 84% win-rate.
Authors: Aakanksha Naik, Bailey Kuehl, Erin Bransom, Doug Downey, Tom Hope
Extracting fine-grained experimental findings from literature can provide massive utility for scientific applications. Prior work has focused on developing annotation schemas and datasets for limited aspects of this problem, leading to simpler information extraction datasets which do not capture the real-world complexity and nuance required for this task. Focusing on biomedicine, this work presents CARE (Clinical Aggregation-oriented Result Extraction) -- a new IE dataset for the task of extracting clinical findings. We develop a new annotation schema capturing fine-grained findings as n-ary relations between entities and attributes, which includes phenomena challenging for current IE systems such as discontinuous entity spans, nested relations, and variable arity n-ary relations. Using this schema, we collect extensive annotations for 700 abstracts from two sources: clinical trials and case reports. We also benchmark the performance of various state-of-the-art IE systems on our dataset, including extractive models and generative LLMs in fully supervised and limited data settings. Our results demonstrate the difficulty of our dataset -- even SOTA models such as GPT4 struggle, particularly on relation extraction. We release our annotation schema and CARE to encourage further research on extracting and aggregating scientific findings from literature.
Authors: Yuhan Liu, Shangbin Feng, Xiaochuang Han, Vidhisha Balachandran, Chan Young Park, Sachin Kumar, Yulia Tsvetkov
In this work, we take a first step towards designing summarization systems that are faithful to the author's opinions and perspectives. Focusing on a case study of preserving political perspectives in news summarization, we find that existing approaches alter the political opinions and stances of news articles in more than 50% of summaries, misrepresenting the intent and perspectives of the news authors. We thus propose P^3Sum, a diffusion model-based summarization approach controlled by political perspective classifiers. In P^3Sum, the political leaning of a generated summary is iteratively evaluated at each decoding step, and any drift from the article's original stance incurs a loss back-propagated to the embedding layers, steering the political stance of the summary at inference time. Extensive experiments on three news summarization datasets demonstrate that P^3Sum outperforms state-of-the-art summarization systems and large language models by up to 11.4% in terms of the success rate of stance preservation, with on-par performance on standard summarization utility metrics. These findings highlight the lacunae that even for state-of-the-art models it is still challenging to preserve author perspectives in news summarization, while P^3Sum presents an important first step towards evaluating and developing summarization systems that are faithful to author intent and perspectives.
Authors: Negar Mokhberian, Myrl G. Marmarelis, Frederic R. Hopp, Valerio Basile, Fred Morstatter, Kristina Lerman
In most classification models, it has been assumed to have a single ground truth label for each data point. However, subjective tasks like toxicity classification can lead to genuine disagreement among annotators. In these cases aggregating labels will result in biased labeling and, consequently, biased models that can overlook minority opinions. Previous studies have shed light on the pitfalls of label aggregation and have introduced a handful of practical approaches to tackle this issue. Recently proposed multi-annotator models, which predict labels individually per annotator, are vulnerable to under-determination for annotators with small samples. This problem is especially the case in crowd-sourced datasets. In this work, we propose Annotator Aware Representations for Texts (AART) for subjective classification tasks. We will show the improvement of our method on metrics that assess the performance on capturing annotators' perspectives. Additionally, our approach involves learning representations for annotators, allowing for an exploration of the captured annotation behaviors.
Authors: Eren Unlu, Unver Ciftci
Due to the limited availability of high quality datasets for training sentence embeddings in Turkish, we propose a training methodology and a regimen to develop a sentence embedding model. The central idea is simple but effective : is to fine-tune a pretrained encoder-decoder model in two consecutive stages, where the first stage involves aligning the embedding space with translation pairs. Thanks to this alignment, the prowess of the main model can be better projected onto the target language in a sentence embedding setting where it can be fine-tuned with high accuracy in short duration with limited target language dataset.
Authors: Miles Williams, Nikolaos Aletras
Pruning and quantization form the foundation of model compression for neural networks, enabling efficient inference for large language models (LLMs). Recently, various quantization and pruning techniques have demonstrated state-of-the-art performance in a post-training setting. They rely upon calibration data, a small set of unlabeled examples, to generate layer activations. However, no prior work has systematically investigated how the calibration data impacts the effectiveness of model compression methods. In this paper, we present the first extensive empirical study on the effect of calibration data upon LLM performance. We trial a variety of pruning and quantization methods, tasks, models, and datasets. Surprisingly, we find substantial variations in downstream task performance, contrasting existing work that suggests a greater level of robustness to the calibration data. Finally, we make a series of recommendations for the effective use of calibration data in LLM quantization and pruning.
Authors: Jiaju Chen, Yuxuan Lu, Shao Zhang, Bingsheng Yao, Yuanzhe Dong, Ying Xu, Yunyao Li, Qianwen Wang, Dakuo Wang, Yuling Sun
AI models (including LLM) often rely on narrative question-answering (QA) datasets to provide customized QA functionalities to support downstream children education applications; however, existing datasets only include QA pairs that are grounded within the given storybook content, but children can learn more when teachers refer the storybook content to real-world knowledge (e.g., commonsense knowledge). We introduce the FairytaleCQA dataset, which is annotated by children education experts, to supplement 278 storybook narratives with educationally appropriate commonsense knowledge. The dataset has 5,868 QA pairs that not only originate from the storybook narrative but also contain the commonsense knowledge grounded by an external knowledge graph (i.e., ConceptNet). A follow-up experiment shows that a smaller model (T5-large) fine-tuned with FairytaleCQA reliably outperforms much larger prompt-engineered LLM (e.g., GPT-4) in this new QA-pair generation task (QAG). This result suggests that: 1) our dataset brings novel challenges to existing LLMs, and 2) human experts' data annotation are still critical as they have much nuanced knowledge that LLMs do not know in the children educational domain.
Authors: Chia-Hsuan Lee, Hao Cheng, Mari Ostendorf
Large language models (LLMs) have revolutionized the landscape of Natural Language Processing systems, but are computationally expensive. To reduce the cost without sacrificing performance, previous studies have explored various approaches to harness the potential of Small Language Models (SLMs) as cost-effective alternatives to their larger counterparts. Driven by findings that SLMs and LLMs exhibit complementary strengths in a structured knowledge extraction task, this work presents a novel SLM/LLM routing framework designed to improve computational efficiency and enhance task performance. First, exemplar pools are created to represent the types of contexts where each LM provides a more reliable answer, leveraging a sentence embedding fine-tuned so that context similarity is close to dialogue state similarity. Then, during inference, the k-nearest exemplars to the testing instance are retrieved, and the instance is routed according to majority vote. In dialogue state tracking tasks, the proposed routing framework enhances performance substantially compared to relying solely on LLMs, while reducing the computational costs by over 50%.
Authors: Chadi Helwe, Tom Calamai, Pierre-Henri Paris, Chloé Clavel, Fabian Suchanek
Fallacies can be used to spread disinformation, fake news, and propaganda, underlining the importance of their detection. Automated detection and classification of fallacies, however, remain challenging, mainly because of the innate subjectivity of the task and the need for a comprehensive, unified approach in existing research. Addressing these limitations, our study introduces a novel taxonomy of fallacies that aligns and refines previous classifications, a new annotation scheme tailored for subjective NLP tasks, and a new evaluation method designed to handle subjectivity, adapted to precision, recall, and F1-Score metrics. Using our annotation scheme, the paper introduces MAFALDA (Multi-level Annotated FALlacy DAtaset), a gold standard dataset. MAFALDA is based on examples from various previously existing fallacy datasets under our unified taxonomy across three levels of granularity. We then evaluate several language models under a zero-shot learning setting using MAFALDA to assess their fallacy detection and classification capability. Our comprehensive evaluation not only benchmarks the performance of these models but also provides valuable insights into their strengths and limitations in addressing fallacious reasoning.
Authors: Jinyoung Park, Ameen Patel, Omar Zia Khan, Hyunwoo J. Kim, Joo-Kyung Kim
Chain-of-Thought (CoT) prompting has boosted the multi-step reasoning capabilities of Large Language Models (LLMs) by generating a series of rationales before the final answer. We analyze the reasoning paths generated by CoT and find two issues in multi-step reasoning: (i) Generating rationales irrelevant to the question, (ii) Unable to compose subquestions or queries for generating/retrieving all the relevant information. To address them, we propose a graph-guided CoT prompting method, which guides the LLMs to reach the correct answer with graph representation/verification steps. Specifically, we first leverage LLMs to construct a "question/rationale graph" by using knowledge extraction prompting given the initial question and the rationales generated in the previous steps. Then, the graph verification step diagnoses the current rationale triplet by comparing it with the existing question/rationale graph to filter out irrelevant rationales and generate follow-up questions to obtain relevant information. Additionally, we generate CoT paths that exclude the extracted graph information to represent the context information missed from the graph extraction. Our graph-guided reasoning method shows superior performance compared to previous CoT prompting and the variants on multi-hop question answering benchmark datasets.
Authors: Wenjie Mo, Jiashu Xu, Qin Liu, Jiongxiao Wang, Jun Yan, Chaowei Xiao, Muhao Chen
Existing studies in backdoor defense have predominantly focused on the training phase, overlooking the critical aspect of testing time defense. This gap becomes particularly pronounced in the context of Large Language Models (LLMs) deployed as Web Services, which typically offer only black-box access, rendering training-time defenses impractical. To bridge this gap, our work introduces defensive demonstrations, an innovative backdoor defense strategy for blackbox large language models. Our method involves identifying the task and retrieving task-relevant demonstrations from an uncontaminated pool. These demonstrations are then combined with user queries and presented to the model during testing, without requiring any modifications/tuning to the black-box model or insights into its internal mechanisms. Defensive demonstrations are designed to counteract the adverse effects of triggers, aiming to recalibrate and correct the behavior of poisoned models during test-time evaluations. Extensive experiments show that defensive demonstrations are effective in defending both instance-level and instruction-level backdoor attacks, not only rectifying the behavior of poisoned models but also surpassing existing baselines in most scenarios.
Authors: Yiqi Liu, Nafise Sadat Moosavi, Chenghua Lin
Automatic evaluation of generated textual content presents an ongoing challenge within the field of NLP. Given the impressive capabilities of modern language models (LMs) across diverse NLP tasks, there is a growing trend to employ these models in creating innovative evaluation metrics for automated assessment of generation tasks. This paper investigates a pivotal question: Do language model-driven evaluation metrics inherently exhibit bias favoring texts generated by the same underlying language model? Specifically, we assess whether prominent LM-based evaluation metrics--namely, BARTScore, T5Score, and GPTScore--demonstrate a favorable bias toward their respective underlying LMs in the context of summarization tasks. Our findings unveil a latent bias, particularly pronounced when such evaluation metrics are used in an reference-free manner without leveraging gold summaries. These results underscore that assessments provided by generative evaluation models can be influenced by factors beyond the inherent text quality, highlighting the necessity of developing more dependable evaluation protocols in the future.
Authors: Yuhan Sun, Mukai Li, Yixin Cao, Kun Wang, Wenxiao Wang, Xingyu Zeng, Rui Zhao
As the use of large language models becomes more widespread, techniques like parameter-efficient fine-tuning and other methods for controlled generation are gaining traction for customizing models and managing their outputs. However, the challenge of precisely controlling how prompts influence these models is an area ripe for further investigation. In response, we introduce ControlPE (Continuously Controllable Prompt Engineering). ControlPE enables finer adjustments to prompt effects, complementing existing prompt engineering, and effectively controls continuous targets. This approach harnesses the power of LoRA (Low-Rank Adaptation) to create an effect akin to prompt weighting, enabling fine-tuned adjustments to the impact of prompts. Our methodology involves generating specialized datasets for prompt distillation, incorporating these prompts into the LoRA model, and carefully adjusting LoRA merging weight to regulate the influence of prompts. This provides a dynamic and adaptable tool for prompt control. Through our experiments, we have validated the practicality and efficacy of ControlPE. It proves to be a promising solution for control a variety of prompts, ranging from generating short responses prompts, refusal prompts to chain-of-thought prompts.
Authors: Junying Chen, Xidong Wang, Anningzhe Gao, Feng Jiang, Shunian Chen, Hongbo Zhang, Dingjie Song, Wenya Xie, Chuyi Kong, Jianquan Li, Xiang Wan, Haizhou Li, Benyou Wang
Adapting a language model into a specific domain, a.k.a `domain adaption', is a common practice when specialized knowledge, e.g. medicine, is not encapsulated in a general language model like Llama2. The challenge lies in the heterogeneity of data across the two training stages, as it varies in languages, genres, or formats. To tackle this and simplify the learning protocol, we propose to transform heterogeneous data, from the both pre-training and supervised stages, into a unified, simple input-output pair format. We validate the new protocol in the domains where proprietary LLMs like ChatGPT perform relatively poorly, such as Traditional Chinese Medicine. The developed model, HuatuoGPT-II, has shown state-of-the-art performance in Chinese medicine domain on a number of benchmarks, e.g. medical licensing exams. It even outperforms proprietary models like ChatGPT and GPT-4 in some aspects, especially in Traditional Chinese Medicine. Expert manual evaluations further validate HuatuoGPT-II's advantages over existing LLMs. Notably, HuatuoGPT-II was benchmarked in a fresh Chinese National Medical Licensing Examination where it achieved the best performance, showcasing not only its effectiveness but also its generalization capabilities.
Authors: Bingsheng Yao, Guiming Chen, Ruishi Zou, Yuxuan Lu, Jiachen Li, Shao Zhang, Sijia Liu, James Hendler, Dakuo Wang
While most existing works on LLM prompt-engineering focus only on how to select a better set of data samples inside one single prompt input (In-Context Learning or ICL), why can't we design and leverage multiple prompt inputs together to further improve the LLM performance? In this work, we propose In-Context Sampling (ICS), a low-resource LLM prompt-engineering technique to produce the most confident prediction results by optimizing the construction of multiple ICL prompt inputs. Extensive experiments with two SOTA LLMs (FlanT5-XL and Mistral-7B) on three NLI datasets (e-SNLI, Multi-NLI, and ANLI) illustrate that ICS can consistently enhance LLM's prediction performance and confidence. An ablation study suggests that a diversity-based ICS strategy may further improve LLM's performance, which sheds light on a new yet promising future research direction.
Authors: Chunyuan Deng, Yilun Zhao, Xiangru Tang, Mark Gerstein, Arman Cohan
Recent observations have underscored a disparity between the inflated benchmark scores and the actual performance of LLMs, raising concerns about potential contamination of evaluation benchmarks. This issue is especially critical for closed-source models and certain open-source models where training data transparency is lacking. In this paper we study data contamination by proposing two methods tailored for both open-source and proprietary LLMs. We first introduce a retrieval-based system to explore potential overlaps between evaluation benchmarks and pretraining corpora. We further present a novel investigation protocol named \textbf{T}estset \textbf{S}lot Guessing (\textit{TS-Guessing}), applicable to both open and proprietary models. This approach entails masking a wrong answer in a multiple-choice question and prompting the model to fill in the gap. Additionally, it involves obscuring an unlikely word in an evaluation example and asking the model to produce it. We find that certain commercial LLMs could surprisingly guess the missing option in various test sets. Specifically, in the TruthfulQA benchmark, we find that LLMs exhibit notable performance improvement when provided with additional metadata in the benchmark. Further, in the MMLU benchmark, ChatGPT and GPT-4 demonstrated an exact match rate of 52\% and 57\%, respectively, in guessing the missing options in benchmark test data. We hope these results underscore the need for more robust evaluation methodologies and benchmarks in the field.
Authors: Nikita Moghe, Patrick Xia, Jacob Andreas, Jason Eisner, Benjamin Van Durme, Harsh Jhamtani
Users of natural language interfaces, generally powered by Large Language Models (LLMs),often must repeat their preferences each time they make a similar request. To alleviate this, we propose including some of a user's preferences and instructions in natural language -- collectively termed standing instructions -- as additional context for such interfaces. For example, when a user states I'm hungry, their previously expressed preference for Persian food will be automatically added to the LLM prompt, so as to influence the search for relevant restaurants. We develop NLSI, a language-to-program dataset consisting of over 2.4K dialogues spanning 17 domains, where each dialogue is paired with a user profile (a set of users specific standing instructions) and corresponding structured representations (API calls). A key challenge in NLSI is to identify which subset of the standing instructions is applicable to a given dialogue. NLSI contains diverse phenomena, from simple preferences to interdependent instructions such as triggering a hotel search whenever the user is booking tickets to an event. We conduct experiments on NLSI using prompting with large language models and various retrieval approaches, achieving a maximum of 44.7% exact match on API prediction. Our results demonstrate the challenges in identifying the relevant standing instructions and their interpretation into API calls.
Authors: Yilun Zhao, Hongjun Liu, Yitao Long, Rui Zhang, Chen Zhao, Arman Cohan
We introduce KnowledgeMath, a novel benchmark designed to evaluate LLMs' capabilities in applying financial knowledge to solve complex math word problems. Compared to prior works, this study features three core advancements. First, KnowledgeMath includes 1,259 problems with a hybrid of textual and tabular content and require college-level knowledge in the finance domain for effective resolution. Second, we provide expert-annotated, detailed solution references in Python program format, ensuring a high-quality benchmark for LLM assessment. Finally, we evaluate a wide spectrum of 14 LLMs with different prompting strategies like Chain-of-Thoughts and Program-of-Thoughts. The current best-performing system (i.e., GPT-4 with Program-of-Thoughts) achieves only 45.4% accuracy, leaving substantial room for improvement. While knowledge-augmented LLMs can improve the performance (e.g., from 23.9% to 32.0% for GPT-3.5), it is still significantly lower the estimated human expert performance of 94%. We believe that KnowledgeMath can facilitate future research on domain-specific knowledge retrieval and augmentation into the math word problem-solving process. We will release the benchmark and code at https://github.com/yale-nlp/KnowledgeMath.
Authors: Shirley Anugrah Hayati, Minhwa Lee, Dheeraj Rajagopal, Dongyeop Kang
Collecting diverse human data on subjective NLP topics is costly and challenging. As Large Language Models (LLMs) have developed human-like capabilities, there is a recent trend in collaborative efforts between humans and LLMs for generating diverse data, offering potential scalable and efficient solutions. However, the extent of LLMs' capability to generate diverse perspectives on subjective topics remains an unexplored question. In this study, we investigate LLMs' capacity for generating diverse perspectives and rationales on subjective topics, such as social norms and argumentative texts. We formulate this problem as diversity extraction in LLMs and propose a criteria-based prompting technique to ground diverse opinions and measure perspective diversity from the generated criteria words. Our results show that measuring semantic diversity through sentence embeddings and distance metrics is not enough to measure perspective diversity. To see how far we can extract diverse perspectives from LLMs, or called diversity coverage, we employ a step-by-step recall prompting for generating more outputs from the model in an iterative manner. As we apply our prompting method to other tasks (hate speech labeling and story continuation), indeed we find that LLMs are able to generate diverse opinions according to the degree of task subjectivity.
Authors: Evgeniia Razumovskaia, Ivan Vulić, Pavle Marković, Tomasz Cichy, Qian Zheng, Tsung-Hsien Wen, Paweł Budzianowski
Factuality is a crucial requirement in information seeking dialogue: the system should respond to the user's queries so that the responses are meaningful and aligned with the knowledge provided to the system. However, most modern large language models suffer from hallucinations, that is, they generate responses not supported by or contradicting the knowledge source. To mitigate the issue and increase faithfulness of information-seeking dialogue systems, we introduce BeInfo, a simple yet effective method that applies behavioural tuning to aid information-seeking dialogue. Relying on three standard datasets, we show that models tuned with BeInfo} become considerably more faithful to the knowledge source both for datasets and domains seen during BeInfo-tuning, as well as on unseen domains, when applied in a zero-shot manner. In addition, we show that the models with 3B parameters (e.g., Flan-T5) tuned with BeInfo demonstrate strong performance on data from real `production' conversations and outperform GPT4 when tuned on a limited amount of such realistic in-domain dialogues.
Authors: Sen Yang, Xin Li, Leyang Cui, Lidong Bing, Wai Lam
Though prompting LLMs with various reasoning structures produces reasoning proofs along with answers, these proofs are not ensured to be causal and reliable due to the inherent defects of LLMs. Tracking such deficiencies, we present a neuro-symbolic integration method, in which a neural LLM is used to represent the knowledge of the problem while an LLM-free symbolic solver is adopted to do deliberative reasoning using the knowledge. Specifically, our customized meta-interpreters allow the production of reasoning proofs and support flexible search strategies. These reasoning proofs are ensured to be causal and reliable because of the deterministic executing nature of the symbolic solvers. Empirically, on ProofWriter, our method surpasses the CoT baseline by nearly double in accuracy and more than triple in proof similarity. On GSM8K, our method also shows accuracy improvements and nearly doubled proof similarity. Our code is released at https://github.com/DAMO-NLP-SG/CaRing
Authors: Yilun Zhao, Yitao Long, Hongjun Liu, Linyong Nan, Lyuhao Chen, Ryo Kamoi, Yixin Liu, Xiangru Tang, Rui Zhang, Arman Cohan
Recent LLMs have demonstrated remarkable performance in solving exam-like math word problems. However, the degree to which these numerical reasoning skills are effective in real-world scenarios, particularly in expert domains, is still largely unexplored. This paper introduces DocMath-Eval, a comprehensive benchmark specifically designed to evaluate the numerical reasoning and problem-solving capabilities of LLMs in the context of understanding and analyzing financial documents containing both text and tables. We evaluate a wide spectrum of 19 LLMs, including those specialized in coding and finance. We also incorporate different prompting strategies (i.e., Chain-of-Thoughts and Program-of-Thoughts) to comprehensively assess the capabilities and limitations of existing LLMs in DocMath-Eval. We found that, although the current best-performing system (i.e., GPT-4), can perform well on simple problems such as calculating the rate of increase in a financial metric within a short document context, it significantly lags behind human experts in more complex problems grounded in longer contexts. We believe DocMath-Eval can be used as a valuable benchmark to evaluate LLMs' capabilities to solve challenging numerical reasoning problems in expert domains. We will release the benchmark and code at https://github.com/yale-nlp/DocMath-Eval.
Authors: Yanzhu Guo, Guokan Shang, Michalis Vazirgiannis, Chloé Clavel
This study investigates the consequences of training large language models (LLMs) on synthetic data generated by their predecessors, an increasingly prevalent practice aimed at addressing the limited supply of human-generated training data. Diverging from the usual emphasis on performance metrics, we focus on the impact of this training methodology on linguistic diversity, especially when conducted recursively over time. To assess this, we developed a set of novel metrics targeting lexical, syntactic, and semantic diversity, applying them in recursive fine-tuning experiments across various natural language generation tasks. Our findings reveal a marked decrease in the diversity of the models' outputs through successive iterations. This trend underscores the potential risks of training LLMs on predecessor-generated text, particularly concerning the preservation of linguistic richness. Our study highlights the need for careful consideration of the long-term effects of such training approaches on the linguistic capabilities of LLMs.
Authors: Iñigo Alonso, Eneko Agirre, Mirella Lapata
Table-to-Text has been traditionally approached as a linear language to text problem. However, visually represented tables are rich in visual information and serve as a concise, effective form of representing data and its relationships. When using text-based approaches, after the linearization process, this information is either lost or represented in a space inefficient manner. This inefficiency has remained a constant challenge for text-based approaches making them struggle with large tables. In this paper, we demonstrate that image representation of tables are more space-efficient than the typical textual linearizations, and multi-modal approaches are competitive in Table-to-Text tasks. We present PixT3, a multimodal table-to-text model that outperforms the state-of-the-art (SotA) in the ToTTo benchmark in a pure Table-to-Text setting while remaining competitive in controlled Table-to-Text scenarios. It also generalizes better in unseen datasets, outperforming ToTTo SotA in all generation settings. Additionally, we introduce a new intermediate training curriculum to reinforce table structural awareness, leading to improved generation and overall faithfulness of the models.
Authors: Maram Hasanain, Fatema Ahmed, Firoj Alam
The use of propagandistic techniques in online communication has increased in recent years, aiming to manipulate online audiences. Efforts to automatically detect and debunk such content have been made, addressing various modeling scenarios. These include determining whether the content (text, image, or multimodal) (i) is propagandistic, (ii) employs one or more techniques, and (iii) includes techniques with identifiable spans. Significant research efforts have been devoted to the first two scenarios compared to the latter. Therefore, in this study, we focus on the task of detecting propagandistic textual spans. We investigate whether large language models such as GPT-4 can be utilized to perform the task of an annotator. For the experiments, we used an in-house developed dataset consisting of annotations from multiple annotators. Our results suggest that providing more information to the model as prompts improves the annotation agreement and performance compared to human annotations. We plan to make the annotated labels from multiple annotators, including GPT-4, available for the community.
Authors: Anirudh Ajith, Sameer Singh, Danish Pruthi
Amidst growing concerns of large language models (LLMs) being misused for generating misinformation or completing homework assignments, watermarking has emerged as an effective solution for distinguishing human-written and LLM-generated text. A prominent watermarking strategy is to embed a signal into generated text by upsampling a (pseudorandomly-chosen) subset of tokens at every generation step. Although this signal is imperceptible to a human reader, it is detectable through statistical testing. However, implanting such signals alters the model's output distribution and can have unintended effects when watermarked LLMs are used for downstream applications. In this work, we evaluate the performance of watermarked LLMs on a diverse suite of tasks, including text classification, textual entailment, reasoning, question answering, translation, summarization, and language modeling. We find that watermarking has negligible impact on the performance of tasks posed as k-class classification problems in the average case. However, the accuracy can plummet to that of a random classifier for some scenarios (that occur with non-negligible probability). Tasks that are cast as multiple-choice questions and short-form generation are surprisingly unaffected by watermarking. For long-form generation tasks, including summarization and translation, we see a drop of 15-20% in the performance due to watermarking. Our findings highlight the trade-offs that users should be cognizant of when using watermarked models, and point to cases where future research could improve existing trade-offs.
Authors: Shicheng Liu, Jialiang Xu, Wesley Tjangnaka, Sina J. Semnani, Chen Jie Yu, Gui Dávid, Monica S. Lam
Many knowledge sources consist of both structured information such as relational databases as well as unstructured free text. Building a conversational interface to such data sources is challenging.
This paper introduces SUQL, Structured and Unstructured Query Language, the first formal executable representation that naturally covers compositions of structured and unstructured data queries. Specifically, it augments SQL with several free-text primitives to form a precise, succinct, and expressive representation. This paper also presents a conversational search agent based on large language models, including a few-shot contextual semantic parser for SUQL.
To validate our approach, we introduce a dataset consisting of crowdsourced questions and conversations about real restaurants. Over 51% of the questions in the dataset require both structured and unstructured data, suggesting that it is a common phenomenon. We show that our few-shot conversational agent based on SUQL finds an entity satisfying all user requirements 89.3% of the time, compared to just 65.0% for a strong and commonly used baseline.
Authors: Qingyu Tan, Hwee Tou Ng, Lidong Bing
Knowledge in the real world is being updated constantly. However, it is costly to frequently update large language models (LLMs). Therefore, it is crucial for LLMs to understand the concept of temporal knowledge. However, prior works on temporal question answering did not emphasize multi-answer and multi-hop types of temporal reasoning. In this paper, we propose a complex temporal question-answering (QA) dataset Complex-TR that focuses on multi-answer and multi-hop temporal reasoning. Besides, we also propose a novel data augmentation strategy to improve the complex temporal reasoning capability and robustness of LLMs. We conducted experiments on multiple temporal QA datasets. Experimental results show that our method is able to improve LLMs' performance on temporal QA benchmarks by significant margins.
Authors: Yuxuan Lu, Bingsheng Yao, Shao Zhang, Yun Wang, Peng Zhang, Tun Lu, Toby Jia-Jun Li, Dakuo Wang
Large Language Models (LLMs) have demonstrated considerable advances, and several claims have been made about their exceeding human performance. However, in real-world tasks, domain knowledge is often required. Low-resource learning methods like Active Learning (AL) have been proposed to tackle the cost of domain expert annotation, raising this question: Can LLMs surpass compact models trained with expert annotations in domain-specific tasks? In this work, we conduct an empirical experiment on four datasets from three different domains comparing SOTA LLMs with small models trained on expert annotations with AL. We found that small models can outperform GPT-3.5 with a few hundreds of labeled data, and they achieve higher or similar performance with GPT-4 despite that they are hundreds time smaller. Based on these findings, we posit that LLM predictions can be used as a warmup method in real-world applications and human experts remain indispensable in tasks involving data annotation driven by domain-specific knowledge.
Authors: Nan Xu, Fei Wang, Ben Zhou, Bang Zheng Li, Chaowei Xiao, Muhao Chen
While large language models (LLMs) have demonstrated increasing power, they have also given rise to a wide range of harmful behaviors. As representatives, jailbreak attacks can provoke harmful or unethical responses from LLMs, even after safety alignment. In this paper, we investigate a novel category of jailbreak attacks specifically designed to target the cognitive structure and processes of LLMs. Specifically, we analyze the safety vulnerability of LLMs in the face of (1) multilingual cognitive overload, (2) veiled expression, and (3) effect-to-cause reasoning. Different from previous jailbreak attacks, our proposed cognitive overload is a black-box attack with no need for knowledge of model architecture or access to model weights. Experiments conducted on AdvBench and MasterKey reveal that various LLMs, including both popular open-source model Llama 2 and the proprietary model ChatGPT, can be compromised through cognitive overload. Motivated by cognitive psychology work on managing cognitive load, we further investigate defending cognitive overload attack from two perspectives. Empirical studies show that our cognitive overload from three perspectives can jailbreak all studied LLMs successfully, while existing defense strategies can hardly mitigate the caused malicious uses effectively.
Authors: Jiayi Wang, David Ifeoluwa Adelani, Sweta Agrawal, Ricardo Rei, Eleftheria Briakou, Marine Carpuat, Marek Masiak, Xuanli He, Sofia Bourhim, Andiswa Bukula, Muhidin Mohamed, Temitayo Olatoye, Hamam Mokayede, Christine Mwase, Wangui Kimotho, Foutse Yuehgoh, Anuoluwapo Aremu, Jessica Ojo, Shamsuddeen Hassan Muhammad, Salomey Osei, Abdul-Hakeem Omotayo, Chiamaka Chukwuneke, Perez Ogayo, Oumaima Hourrane, Salma El Anigri, Lolwethu Ndolela, Thabiso Mangwana, Shafie Abdi Mohamed, Ayinde Hassan, Oluwabusayo Olufunke Awoyomi, Lama Alkhaled, Sana Al-Azzawi, Naome A. Etori, Millicent Ochieng, Clemencia Siro, Samuel Njoroge, Eric Muchiri, Wangari Kimotho, Lyse Naomi Wamba Momo, Daud Abolade, Simbiat Ajao, Tosin Adewumi, Iyanuoluwa Shode, Ricky Macharm, Ruqayya Nasir Iro, Saheed S. Abdullahi, Stephen E. Moore, et al. (10 additional authors not shown)
Despite the progress we have recorded in scaling multilingual machine translation (MT) models and evaluation data to several under-resourced African languages, it is difficult to measure accurately the progress we have made on these languages because evaluation is often performed on n-gram matching metrics like BLEU that often have worse correlation with human judgments. Embedding-based metrics such as COMET correlate better; however, lack of evaluation data with human ratings for under-resourced languages, complexity of annotation guidelines like Multidimensional Quality Metrics (MQM), and limited language coverage of multilingual encoders have hampered their applicability to African languages. In this paper, we address these challenges by creating high-quality human evaluation data with a simplified MQM guideline for error-span annotation and direct assessment (DA) scoring for 13 typologically diverse African languages. Furthermore, we develop AfriCOMET, a COMET evaluation metric for African languages by leveraging DA training data from high-resource languages and African-centric multilingual encoder (AfroXLM-Roberta) to create the state-of-the-art evaluation metric for African languages MT with respect to Spearman-rank correlation with human judgments (+0.406).
Authors: Yimin Jing, Renren Jin, Jiahao Hu, Huishi Qiu, Xiaohua Wang, Peng Wang, Deyi Xiong
The effective assessment of the instruction-following ability of large language models (LLMs) is of paramount importance. A model that cannot adhere to human instructions might be not able to provide reliable and helpful responses. In pursuit of this goal, various benchmarks have been constructed to evaluate the instruction-following capacity of these models. However, these benchmarks are limited to a single language and are constructed using automated approaches, which restricts their applicability and the quality of the test examples they contain. To bridge this gap, we introduce the FollowEval benchmark in this paper. This benchmark is composed of instances in both English and Chinese, and all test examples are crafted by human experts. Furthermore, the FollowEval benchmark is designed to assess LLMs across five critical dimensions of instruction following: string manipulation, commonsense reasoning, logical reasoning, spatial reasoning, and response constraints. To enhance the complexity and present a sufficient challenge, each test example is designed to evaluate more than one dimension. We have evaluated various LLMs using the FollowEval benchmark and found that their performance significantly lags behind that of humans. This highlights the considerable room for improvement in the instruction-following ability of these models.
Authors: Katharina Stein, Alexander Koller
LLMs are being increasingly used for planning-style tasks, but their capabilities for planning and reasoning are poorly understood. We present a novel method for automatically converting planning benchmarks written in PDDL into textual descriptions and offer a benchmark dataset created with our method. We show that while the best LLM planners do well on many planning tasks, others remain out of reach of current methods.
Authors: Liang Chen, Yatao Bian, Yang Deng, Shuaiyi Li, Bingzhe Wu, Peilin Zhao, Kam-fai Wong
Text watermarking has emerged as an important technique for detecting machine-generated text. However, existing methods can severely degrade text quality due to arbitrary vocabulary partitioning, which disrupts the language model's expressiveness and impedes textual coherence. To mitigate this, we introduce XMark, a novel approach that capitalizes on text redundancy within the lexical space. Specifically, XMark incorporates a mutually exclusive rule for synonyms during the language model decoding process, thereby integrating prior knowledge into vocabulary partitioning and preserving the capabilities of language generation. We present theoretical analyses and empirical evidence demonstrating that XMark substantially enhances text generation fluency while maintaining watermark detectability. Furthermore, we investigate watermarking's impact on the emergent abilities of large language models, including zero-shot and few-shot knowledge recall, logical reasoning, and instruction following. Our comprehensive experiments confirm that XMark consistently outperforms existing methods in retaining these crucial capabilities of LLMs.
Authors: Sarah Masud, Mohammad Aflah Khan, Md. Shad Akhtar, Tanmoy Chakraborty
As hate speech continues to proliferate on the web, it is becoming increasingly important to develop computational methods to mitigate it. Reactively, using black-box models to identify hateful content can perplex users as to why their posts were automatically flagged as hateful. On the other hand, proactive mitigation can be achieved by suggesting rephrasing before a post is made public. However, both mitigation techniques require information about which part of a post contains the hateful aspect, i.e., what spans within a text are responsible for conveying hate. Better detection of such spans can significantly reduce explicitly hateful content on the web. To further contribute to this research area, we organized HateNorm at HASOC-FIRE 2023, focusing on explicit span detection in English Tweets. A total of 12 teams participated in the competition, with the highest macro-F1 observed at 0.58.
Authors: Yuliang Liu, Xiangru Tang, Zefan Cai, Junjie Lu, Yichi Zhang, Yanjun Shao, Zexuan Deng, Helan Hu, Zengxian Yang, Kaikai An, Ruijun Huang, Shuzheng Si, Sheng Chen, Haozhe Zhao, Zhengliang Li, Liang Chen, Yiming Zong, Yan Wang, Tianyu Liu, Zhiwei Jiang, Baobao Chang, Yujia Qin, Wangchunshu Zhou, Yilun Zhao, Arman Cohan, Mark Gerstein
Large language models have shown promising performance in code generation benchmarks. However, a considerable divide exists between these benchmark achievements and their practical applicability, primarily attributed to real-world programming's reliance on pre-existing libraries. Instead of evaluating LLMs to code from scratch, this work aims to propose a new evaluation setup where LLMs use open-source libraries to finish machine learning tasks. Therefore, we propose ML-Bench, an expansive benchmark developed to assess the effectiveness of LLMs in leveraging existing functions in open-source libraries. Consisting of 10044 samples spanning 130 tasks over 14 notable machine learning GitHub repositories. In this setting, given a specific machine learning task instruction and the accompanying README in a codebase, an LLM is tasked to generate code to accomplish the task. This necessitates the comprehension of long and language-code interleaved documents, as well as the understanding of complex cross-file code structures, introducing new challenges. Notably, while GPT-4 exhibits remarkable improvement over other LLMs, it manages to accomplish only 39.73\% of the tasks, leaving a huge space for improvement. We address these challenges by proposing ML-Agent, designed to effectively navigate the codebase, locate documentation, retrieve code, and generate executable code. Empirical results demonstrate that ML-Agent, built upon GPT-4, results in further improvements. Code, data, and models are available at \url{https://ml-bench.github.io/}.
Authors: Joseph J. Peper, Wenzhao Qiu, Lu Wang
We investigate pre-training techniques for abstractive multi-document summarization (MDS), which is much less studied than summarizing single documents. Though recent work has demonstrated the effectiveness of highlighting information salience for pre-training strategy design, it struggles to generate abstractive and reflective summaries, which are critical properties for MDS. To this end, we present PELMS, a pre-trained model that uses objectives based on semantic coherence heuristics and faithfulness constraints with un-labeled multi-document inputs, to promote the generation of concise, fluent, and faithful summaries. To support the training of PELMS, we compile MultiPT, a multi-document pre-training corpus containing over 93 million documents to form more than 3 million unlabeled topic-centric document clusters, covering diverse genres such as product reviews, news, and general knowledge. We perform extensive evaluation of PELMS in low-shot settings on a wide range of MDS datasets. Our approach consistently outperforms competitive comparisons with respect to overall informativeness, abstractiveness, coherence, and faithfulness.
Authors: Tilahun Abedissa Taffa, Ricardo Usbeck
This paper presents a scholarly Knowledge Graph Question Answering (KGQA) that answers bibliographic natural language questions by leveraging a large language model (LLM) in a few-shot manner. The model initially identifies the top-n similar training questions related to a given test question via a BERT-based sentence encoder and retrieves their corresponding SPARQL. Using the top-n similar question-SPARQL pairs as an example and the test question creates a prompt. Then pass the prompt to the LLM and generate a SPARQL. Finally, runs the SPARQL against the underlying KG - ORKG (Open Research KG) endpoint and returns an answer. Our system achieves an F1 score of 99.0%, on SciQA - one of the Scholarly-QALD-23 challenge benchmarks.
Authors: Wolfgang Otto, Matthäus Zloch, Lu Gan, Saurav Karmakar, Stefan Dietze
Named Entity Recognition (NER) models play a crucial role in various NLP tasks, including information extraction (IE) and text understanding. In academic writing, references to machine learning models and datasets are fundamental components of various computer science publications and necessitate accurate models for identification. Despite the advancements in NER, existing ground truth datasets do not treat fine-grained types like ML model and model architecture as separate entity types, and consequently, baseline models cannot recognize them as such. In this paper, we release a corpus of 100 manually annotated full-text scientific publications and a first baseline model for 10 entity types centered around ML models and datasets. In order to provide a nuanced understanding of how ML models and datasets are mentioned and utilized, our dataset also contains annotations for informal mentions like "our BERT-based model" or "an image CNN". You can find the ground truth dataset and code to replicate model training at https://data.gesis.org/gsap/gsap-ner.
Authors: Junlei Zhang, Hongliang He, Nirui Song, Shuyuan He, \\Shuai Zhang, Huachuan Qiu, Anqi Li, Lizhi Ma, Zhenzhong Lan
As Large Language Models (LLMs) are becoming prevalent in various fields, there is an urgent need for improved NLP benchmarks that encompass all the necessary knowledge of individual discipline. Many contemporary benchmarks for foundational models emphasize a broad range of subjects but often fall short in presenting all the critical subjects and encompassing necessary professional knowledge of them. This shortfall has led to skewed results, given that LLMs exhibit varying performance across different subjects and knowledge areas. To address this issue, we present psybench, the first comprehensive Chinese evaluation suite that covers all the necessary knowledge required for graduate entrance exams. psybench offers a deep evaluation of a model's strengths and weaknesses in psychology through multiple-choice questions. Our findings show significant differences in performance across different sections of a subject, highlighting the risk of skewed results when the knowledge in test sets is not balanced. Notably, only the ChatGPT model reaches an average accuracy above $70\%$, indicating that there is still plenty of room for improvement. We expect that psybench will help to conduct thorough evaluations of base models' strengths and weaknesses and assist in practical application in the field of psychology.
Authors: Debarati Das, Ishaan Gupta, Jaideep Srivastava, Dongyeop Kang
Large language models (LLMs) are revolutionizing various fields by leveraging large text corpora for context-aware intelligence. Due to the context size, however, encoding an entire graph with LLMs is fundamentally limited. This paper explores how to better integrate graph data with LLMs and presents a novel approach using various encoding modalities (e.g., text, image, and motif) and approximation of global connectivity of a graph using different prompting methods to enhance LLMs' effectiveness in handling complex graph structures. The study also introduces GraphTMI, a new benchmark for evaluating LLMs in graph structure analysis, focusing on factors such as homophily, motif presence, and graph difficulty. Key findings reveal that image modality, supported by advanced vision-language models like GPT-4V, is more effective than text in managing token limits while retaining critical information. The research also examines the influence of different factors on each encoding modality's performance. This study highlights the current limitations and charts future directions for LLMs in graph understanding and reasoning tasks.
Authors: Ziyi Ye, Qingyao Ai, Yiqun Liu, Min Zhang, Christina Lioma, Tuukka Ruotsalo
Generating human language through non-invasive brain-computer interfaces (BCIs) has the potential to unlock many applications, such as serving disabled patients and improving communication. Currently, however, generating language via BCIs has been previously successful only within a classification setup for selecting pre-generated sentence continuation candidates with the most likely cortical semantic representation. Inspired by recent research that revealed associations between the brain and the large computational language models, we propose a generative language BCI that utilizes the capacity of a large language model (LLM) jointly with a semantic brain decoder to directly generate language from functional magnetic resonance imaging (fMRI) input. The proposed model can generate coherent language sequences aligned with the semantic content of visual or auditory language stimuli perceived, without prior knowledge of any pre-generated candidates. We compare the language generated from the presented model with a random control, pre-generated language selection approach, and a standard LLM, which generates common coherent text solely based on the next word likelihood according to statistical language training data. The proposed model is found to generate language that is more aligned with semantic stimulus in response to which brain input is sampled. Our findings demonstrate the potential and feasibility of employing BCIs in direct language generation.
Authors: Qirui Tang, Wenkang Jiang, Yihua Du, Lei Lin
In social media networks, users produce a large amount of text content anytime, providing researchers with a valuable approach to digging for personality-related information. Personality detection based on user-generated texts is a universal method that can be used to build user portraits. The presence of noise in social media texts hinders personality detection. However, previous studies have not fully addressed this challenge. Inspired by the scanning reading technique, we propose an attention-based information extraction mechanism (AIEM) for long texts, which is applied to quickly locate valuable pieces of information, and focus more attention on the deep semantics of key pieces. Then, we provide a novel attention-based denoising framework (ADF) for personality detection tasks and achieve state-of-the-art performance on two commonly used datasets. Notably, we obtain an average accuracy improvement of 10.2% on the gold standard Twitter-Myers-Briggs Type Indicator (Twitter-MBTI) dataset. We made our code publicly available on GitHub. We shed light on how AIEM works to magnify personality-related signals.
Authors: Yao Qiang, Xiangyu Zhou, Dongxiao Zhu
In-context learning (ICL) has emerged as a powerful paradigm leveraging LLMs for specific tasks by utilizing labeled examples as demonstrations in the precondition prompts. Despite its promising performance, ICL suffers from instability with the choice and arrangement of examples. Additionally, crafted adversarial attacks pose a notable threat to the robustness of ICL. However, existing attacks are either easy to detect, rely on external models, or lack specificity towards ICL. To address these issues, this work introduces a novel transferable attack for ICL, aiming to hijack LLMs to generate the targeted response. The proposed LLM hijacking attack leverages a gradient-based prompt search method to learn and append imperceptible adversarial suffixes to the in-context demonstrations. Extensive experimental results on various tasks and datasets demonstrate the effectiveness of our LLM hijacking attack, resulting in a distracted attention towards adversarial tokens, consequently leading to the targeted unwanted outputs.
Authors: Tomoyuki Yamakami
Nonuniform families of polynomial-size finite automata, which are series of indexed finite automata having polynomially many inner states, are used in the past literature to solve nonuniform families of promise decision problems. Among such nonuniform families of finite automata, we focus our attention, in particular, on the variants of nondeterministic finite automata, which have at most "one" (unambiguous), "polynomially many" (few) accepting computation paths, or unambiguous/few computation paths leading to each fixed configuration. When such machines are limited to make only one-way head moves, we can prove with no unproven hardness assumptions that some of these variants are different in computational power from each other. As for two-way machines restricted to instances of polynomially-bounded length, families of two-way polynomial-size nondeterministic finite automata are equivalent in power to families of polynomial-size unambiguous finite automata.
Authors: Sagi Pendzel, Tomer Wullach, Amir Adler, Einat Minkov
Automatic hate speech detection using deep neural models is hampered by the scarcity of labeled datasets, leading to poor generalization. To mitigate this problem, generative AI has been utilized to generate large amounts of synthetic hate speech sequences from available labeled examples, leveraging the generated data in finetuning large pre-trained language models (LLMs). In this chapter, we provide a review of relevant methods, experimental setups and evaluation of this approach. In addition to general LLMs, such as BERT, RoBERTa and ALBERT, we apply and evaluate the impact of train set augmentation with generated data using LLMs that have been already adapted for hate detection, including RoBERTa-Toxicity, HateBERT, HateXplain, ToxDect, and ToxiGen. An empirical study corroborates our previous findings, showing that this approach improves hate speech generalization, boosting recall performance across data distributions. In addition, we explore and compare the performance of the finetuned LLMs with zero-shot hate detection using a GPT-3.5 model. Our results demonstrate that while better generalization is achieved using the GPT-3.5 model, it achieves mediocre recall and low precision on most datasets. It is an open question whether the sensitivity of models such as GPT-3.5, and onward, can be improved using similar techniques of text generation.
Authors: Sergey Slavnov
It is known that different categorial grammars have surface representation in a fragment of first order multiplicative linear logic (MLL1). We show that the fragment of interest is equivalent to the recently introduced extended tensor type calculus (ETTC). ETTC is a calculus of specific typed terms, which represent tuples of strings, more precisely bipartite graphs decorated with strings. Types are derived from linear logic formulas, and rules correspond to concrete operations on these string-labeled graphs, so that they can be conveniently visualized. This provides the above mentioned fragment of MLL1 that is relevant for language modeling not only with some alternative syntax and intuitive geometric representation, but also with an intrinsic deductive system, which has been absent.
In this work we consider a non-trivial notationally enriched variation of the previously introduced ETTC, which allows more concise and transparent computations. We present both a cut-free sequent calculus and a natural deduction formalism.
Authors: Angelica Chen, David M. Dohan, David R. So
Given the recent impressive accomplishments of language models (LMs) for code generation, we explore the use of LMs as adaptive mutation and crossover operators for an evolutionary neural architecture search (NAS) algorithm. While NAS still proves too difficult a task for LMs to succeed at solely through prompting, we find that the combination of evolutionary prompt engineering with soft prompt-tuning, a method we term EvoPrompting, consistently finds diverse and high performing models. We first demonstrate that EvoPrompting is effective on the computationally efficient MNIST-1D dataset, where EvoPrompting produces convolutional architecture variants that outperform both those designed by human experts and naive few-shot prompting in terms of accuracy and model size. We then apply our method to searching for graph neural networks on the CLRS Algorithmic Reasoning Benchmark, where EvoPrompting is able to design novel architectures that outperform current state-of-the-art models on 21 out of 30 algorithmic reasoning tasks while maintaining similar model size. EvoPrompting is successful at designing accurate and efficient neural network architectures across a variety of machine learning tasks, while also being general enough for easy adaptation to other tasks beyond neural network design.
Authors: Tilahun Abedissa, Ricardo Usbeck, Yaregal Assabie
Question Answering (QA) returns concise answers or answer lists from natural language text given a context document. Many resources go into curating QA datasets to advance robust models' development. There is a surge of QA datasets for languages like English, however, this is not true for Amharic. Amharic, the official language of Ethiopia, is the second most spoken Semitic language in the world. There is no published or publicly available Amharic QA dataset. Hence, to foster the research in Amharic QA, we present the first Amharic QA (AmQA) dataset. We crowdsourced 2628 question-answer pairs over 378 Wikipedia articles. Additionally, we run an XLMR Large-based baseline model to spark open-domain QA research interest. The best-performing baseline achieves an F-score of 69.58 and 71.74 in reader-retriever QA and reading comprehension settings respectively.
Authors: Yixin Liu, Alexander R. Fabbri, Yilun Zhao, Pengfei Liu, Shafiq Joty, Chien-Sheng Wu, Caiming Xiong, Dragomir Radev
Interpretability and efficiency are two important considerations for the adoption of neural automatic metrics. In this work, we develop strong-performing automatic metrics for reference-based summarization evaluation, based on a two-stage evaluation pipeline that first extracts basic information units from one text sequence and then checks the extracted units in another sequence. The metrics we developed include two-stage metrics that can provide high interpretability at both the fine-grained unit level and summary level, and one-stage metrics that achieve a balance between efficiency and interpretability. We make the developed tools publicly available at https://github.com/Yale-LILY/AutoACU.
Authors: Honghua Zhang, Meihua Dang, Nanyun Peng, Guy Van den Broeck
Despite the success of autoregressive large language models in text generation, it remains a major challenge to generate text that satisfies complex constraints: sampling from the conditional distribution ${\Pr}(\text{text} | \alpha)$ is intractable for even the simplest lexical constraints $\alpha$. To overcome this challenge, we propose to use tractable probabilistic models (TPMs) to impose lexical constraints in autoregressive text generation models, which we refer to as GeLaTo (Generating Language with Tractable Constraints). To demonstrate the effectiveness of this framework, we use distilled hidden Markov models, where we can efficiently compute ${\Pr}(\text{text} | \alpha)$, to guide autoregressive generation from GPT2. GeLaTo achieves state-of-the-art performance on challenging benchmarks for constrained text generation (e.g., CommonGen), beating various strong baselines by a large margin. Our work not only opens up new avenues for controlling large language models but also motivates the development of more expressive TPMs.
Authors: Hang Jiang, Xiajie Zhang, Xubo Cao, Jad Kabbara
Despite the many use cases for large language models (LLMs) in creating personalized chatbots, there has been limited research on evaluating the extent to which the behaviors of personalized LLMs accurately and consistently reflect specific personality traits. We consider studying the behavior of LLM-based agents, referred to as LLM personas, and present a case study with ChatGPT and GPT-4. The study investigates whether LLMs can generate content that aligns with their assigned personality profiles. To this end, we create distinct LLM personas based on the Big Five personality model, have them complete the 44-item Big Five Inventory (BFI) personality test and a story writing task, and then assess their essays with automatic and human evaluations. Results show that LLM personas' self-reported BFI scores are consistent with their designated personality types, with large effect sizes observed across five traits. Additionally, there are significant correlations between the assigned personality types and certain psycholinguistic features of their writings, as measured by the Linguistic Inquiry and Word Count (LIWC) tool. Interestingly, human evaluators perceive the stories as less personal when told that the stories are authored by AI. However, their judgments on other aspects of the writing such as readability, cohesiveness, redundancy, likeability, and believability remain largely unaffected. Notably, when evaluators were informed about the AI authorship, their accuracy in identifying the intended personality traits from the stories decreased by more than 10% for some traits. This research marks a significant step forward in understanding the capabilities of LLMs to express personality traits.
Authors: Eve Fleisig, Rediet Abebe, Dan Klein
Though majority vote among annotators is typically used for ground truth labels in natural language processing, annotator disagreement in tasks such as hate speech detection may reflect differences in opinion across groups, not noise. Thus, a crucial problem in hate speech detection is determining whether a statement is offensive to the demographic group that it targets, when that group may constitute a small fraction of the annotator pool. We construct a model that predicts individual annotator ratings on potentially offensive text and combines this information with the predicted target group of the text to model the opinions of target group members. We show gains across a range of metrics, including raising performance over the baseline by 22% at predicting individual annotators' ratings and by 33% at predicting variance among annotators, which provides a metric for model uncertainty downstream. We find that annotator ratings can be predicted using their demographic information and opinions on online content, without the need to track identifying annotator IDs that link each annotator to their ratings. We also find that use of non-invasive survey questions on annotators' online experiences helps to maximize privacy and minimize unnecessary collection of demographic information when predicting annotators' opinions.
Authors: Pranjal Aggarwal, Aman Madaan, Yiming Yang, Mausam
A popular approach for improving the correctness of output from large language models (LLMs) is Self-Consistency - poll the LLM multiple times and output the most frequent solution. Existing Self-Consistency techniques always generate a constant number of samples per question, where a better approach will be to non-uniformly distribute the available budget based on the amount of agreement in the samples generated so far. In response, we introduce Adaptive-Consistency, a cost-efficient, model-agnostic technique that dynamically adjusts the number of samples per question using a lightweight stopping criterion. Our experiments over 17 reasoning and code generation datasets and three LLMs demonstrate that Adaptive-Consistency reduces sample budget by up to 7.9 times with an average accuracy drop of less than 0.1%. Our code and data are available at https://www.sample-step-by-step.info
Authors: Yixin Liu, Kejian Shi, Katherine S He, Longtian Ye, Alexander R. Fabbri, Pengfei Liu, Dragomir Radev, Arman Cohan
Recent studies have found that summaries generated by large language models (LLMs) are favored by human annotators over the original reference summaries in commonly used summarization datasets. Therefore, we investigate a new learning setting of text summarization models that considers the LLMs as the reference or the gold-standard oracle on these datasets. To examine the standard practices that are aligned with this new learning setting, we investigate two LLM-based summary quality evaluation methods for model training and adopt a contrastive learning training method to leverage the LLM-guided learning signals. Our experiments on the CNN/DailyMail and XSum datasets demonstrate that smaller summarization models can achieve similar performance as LLMs under LLM-based evaluation. However, we found that the smaller models can not yet reach LLM-level performance under human evaluation despite promising improvements brought by our proposed training methods. Meanwhile, we perform a meta-analysis on this new learning setting that reveals a discrepancy between human and LLM-based evaluation, highlighting the benefits and risks of this LLM-as-reference setting we investigated.
Authors: Tarek Naous, Michael J. Ryan, Alan Ritter, Wei Xu
It is important that language models appropriately adapt to specific cultural contexts. However, as we show in this paper, multilingual and Arabic monolingual language models default to Western culture even when prompted in Arabic and contextualized by an Arab cultural setting. To measure this Western bias, we introduce CAMeL, a dataset of naturally occurring Arabic prompts spanning eight diverse cultural aspects and an extensive list of 20,504 cultural targets corresponding to Arab or Western culture. Using CAMeL, we show that models favor Western targets and demonstrate cultural unfairness on downstream tasks such as named entity recognition and sentiment analysis. Our analyses of pretraining corpora also reveal that commonly used sources such as Wikipedia may not be suited to build culturally aware models, underscoring the importance of carefully curating pretraining data in constructing language models to serve a global population.
Authors: Shuyang Cao, Lu Wang
Long document summarization systems are critical for domains with lengthy and jargonladen text, yet they present significant challenges to researchers and developers with limited computing resources. Existing solutions mainly focus on efficient attentions or divide-and-conquer strategies. The former reduces theoretical time complexity, but is still memory-heavy. The latter methods sacrifice global context, leading to uninformative and incoherent summaries. This work aims to leverage the memory-efficient nature of divide-and-conquer methods while preserving global context. Concretely, our framework AWESOME uses two novel mechanisms: (1) External memory mechanisms track previously encoded document segments and their corresponding summaries, to enhance global document understanding and summary coherence. (2) Global salient content is further identified beforehand to augment each document segment to support its summarization. Extensive experiments on diverse genres of text, including government reports, transcripts, scientific papers, and novels, show that AWESOME produces summaries with improved informativeness, faithfulness, and coherence than competitive baselines on longer documents, while having a smaller GPU memory footprint.
Authors: Michael J.Q. Zhang, Eunsol Choi
While large language models are able to retain vast amounts of world knowledge seen during pretraining, such knowledge is prone to going out of date and is nontrivial to update. Furthermore, these models are often used under temporal misalignment, tasked with answering questions about the present, despite having only been trained on data collected in the past. To mitigate the effects of temporal misalignment, we propose fact duration prediction: the task of predicting how long a given fact will remain true. In our experiments, we demonstrate that identifying which facts are prone to rapid change can help models avoid reciting outdated information and determine which predictions require seeking out up-to-date knowledge sources. We also show how modeling fact duration improves calibration for knowledge-intensive tasks, such as open-retrieval question answering, under temporal misalignment, by discarding volatile facts. Our data and code are released publicly at https://github.com/mikejqzhang/mitigating_misalignment.
Authors: Noam Rotstein, David Bensaid, Shaked Brody, Roy Ganz, Ron Kimmel
The advent of vision-language pre-training techniques enhanced substantial progress in the development of models for image captioning. However, these models frequently produce generic captions and may omit semantically important image details. This limitation can be traced back to the image-text datasets; while their captions typically offer a general description of image content, they frequently omit salient details. Considering the magnitude of these datasets, manual reannotation is impractical, emphasizing the need for an automated approach. To address this challenge, we leverage existing captions and explore augmenting them with visual details using "frozen" vision experts including an object detector, an attribute recognizer, and an Optical Character Recognizer (OCR). Our proposed method, FuseCap, fuses the outputs of such vision experts with the original captions using a large language model (LLM), yielding comprehensive image descriptions. We automatically curate a training set of 12M image-enriched caption pairs. These pairs undergo extensive evaluation through both quantitative and qualitative analyses. Subsequently, this data is utilized to train a captioning generation BLIP-based model. This model outperforms current state-of-the-art approaches, producing more precise and detailed descriptions, demonstrating the effectiveness of the proposed data-centric approach. We release this large-scale dataset of enriched image-caption pairs for the community.
Authors: Yihan Wang, Jatin Chauhan, Wei Wang, Cho-Jui Hsieh
Despite the demonstrated empirical efficacy of prompt tuning to adapt a pretrained language model for a new task, the theoretical underpinnings of the difference between "tuning parameters before the input" against "the tuning of model weights" are limited. We thus take one of the first steps to understand the role of soft-prompt tuning for transformer-based architectures. By considering a general purpose architecture, we analyze prompt tuning from the lens of both: universal approximation and limitations with finite-depth fixed-weight pretrained transformers for continuous-valued functions. Our universality result guarantees the existence of a strong transformer with a prompt to approximate any sequence-to-sequence function in the set of Lipschitz functions. The limitations of prompt tuning for limited-depth transformers are first proved by constructing a set of datasets, that cannot be memorized by a prompt of any length for a given single encoder layer. We also provide a lower bound on the required number of tunable prompt parameters and compare the result with the number of parameters required for a low-rank update (based on LoRA) for a single-layer setting. We finally extend our analysis to multi-layer settings by providing sufficient conditions under which the transformer can at best learn datasets from invertible functions only. Our theoretical claims are also corroborated by empirical results.
Authors: Jingwei Ni, Julia Bingler, Chiara Colesanti-Senni, Mathias Kraus, Glen Gostlow, Tobias Schimanski, Dominik Stammbach, Saeid Ashraf Vaghefi, Qian Wang, Nicolas Webersinke, Tobias Wekhof, Tingyu Yu, Markus Leippold
This paper introduces a novel approach to enhance Large Language Models (LLMs) with expert knowledge to automate the analysis of corporate sustainability reports by benchmarking them against the Task Force for Climate-Related Financial Disclosures (TCFD) recommendations. Corporate sustainability reports are crucial in assessing organizations' environmental and social risks and impacts. However, analyzing these reports' vast amounts of information makes human analysis often too costly. As a result, only a few entities worldwide have the resources to analyze these reports, which could lead to a lack of transparency. While AI-powered tools can automatically analyze the data, they are prone to inaccuracies as they lack domain-specific expertise. This paper introduces a novel approach to enhance LLMs with expert knowledge to automate the analysis of corporate sustainability reports. We christen our tool CHATREPORT, and apply it in a first use case to assess corporate climate risk disclosures following the TCFD recommendations. CHATREPORT results from collaborating with experts in climate science, finance, economic policy, and computer science, demonstrating how domain experts can be involved in developing AI tools. We make our prompt templates, generated data, and scores available to the public to encourage transparency.
Authors: Georg Wenzel, Adam Jatowt
Temporal commonsense reasoning refers to the ability to understand the typical temporal context of phrases, actions, and events, and use it to reason over problems requiring such knowledge. This trait is essential in temporal natural language processing tasks, with possible applications such as timeline summarization, temporal question answering, and temporal natural language inference. Recent research on the performance of large language models suggests that, although they are adept at generating syntactically correct sentences and solving classification tasks, they often take shortcuts in their reasoning and fall prey to simple linguistic traps. This article provides an overview of research in the domain of temporal commonsense reasoning, particularly focusing on enhancing language model performance through a variety of augmentations and their evaluation across a growing number of datasets. However, these augmented models still struggle to approach human performance on reasoning tasks over temporal common sense properties, such as the typical occurrence times, orderings, or durations of events. We further emphasize the need for careful interpretation of research to guard against overpromising evaluation results in light of the shallow reasoning present in transformers. This can be achieved by appropriately preparing datasets and suitable evaluation metrics.
Authors: Chaoyi Wu, Xiaoman Zhang, Ya Zhang, Yanfeng Wang, Weidi Xie
In this study, we aim to initiate the development of Radiology Foundation Model, termed as RadFM. We consider the construction of foundational models from three perspectives, namely, dataset construction, model design, and thorough evaluation. Our contribution can be concluded as follows: (i), we construct a large-scale Medical Multi-modal Dataset, MedMD, which consists of 16M 2D and 3D medical scans with high-quality text descriptions or reports across various data formats, modalities, and tasks, covering over 5000 distinct diseases. To the best of our knowledge, this is the first large-scale, high-quality, medical visual-language dataset, with both 2D and 3D scans; (ii), we propose an architecture that enables visually conditioned generative pre-training, i.e., allowing for integration of text input with 2D or 3D medical scans, and generate responses for diverse radiologic tasks. The model was initially pre-trained on MedMD and subsequently fine-tuned on the domain-specific dataset, which is a radiologic cleaned version of MedMD, containing 3M radiologic visual-language pairs, termed as RadMD; (iii), we propose a new evaluation benchmark, RadBench, that comprises five tasks, including modality recognition, disease diagnosis, visual question answering, report generation and rationale diagnosis, aiming to comprehensively assess the capability of foundation models in handling practical clinical problems. We conduct both automatic and human evaluation on RadBench, in both cases, RadFM outperforms existing multi-modal foundation models, that are publicaly accessible, including Openflamingo, MedFlamingo, MedVInT and GPT-4V. Additionally, we also adapt RadFM for different public benchmarks, surpassing existing SOTAs on diverse datasets. All codes, data, and model checkpoint will all be made publicly available to promote further research and development in the field.
Authors: Jen-tse Huang, Man Ho Lam, Eric John Li, Shujie Ren, Wenxuan Wang, Wenxiang Jiao, Zhaopeng Tu, Michael R. Lyu
Recently, the community has witnessed the advancement of Large Language Models (LLMs), which have shown remarkable performance on various downstream tasks. Led by powerful models like ChatGPT and Claude, LLMs are revolutionizing how users engage with software, assuming more than mere tools but intelligent assistants. Consequently, evaluating LLMs' anthropomorphic capabilities becomes increasingly important in contemporary discourse. Utilizing the emotion appraisal theory from psychology, we propose to evaluate the empathy ability of LLMs, i.e., how their feelings change when presented with specific situations. After a careful and comprehensive survey, we collect a dataset containing over 400 situations that have proven effective in eliciting the eight emotions central to our study. Categorizing the situations into 36 factors, we conduct a human evaluation involving more than 1,200 subjects worldwide. With the human evaluation results as references, our evaluation includes five LLMs, covering both commercial and open-source models, including variations in model sizes, featuring the latest iterations, such as GPT-4 and LLaMA 2. A conclusion can be drawn from the results that, despite several misalignments, LLMs can generally respond appropriately to certain situations. Nevertheless, they fall short in alignment with the emotional behaviors of human beings and cannot establish connections between similar situations. Our collected dataset of situations, the human evaluation results, and the code of our testing framework, dubbed EmotionBench, is made publicly in https://github.com/CUHK-ARISE/EmotionBench. We aspire to contribute to the advancement of LLMs regarding better alignment with the emotional behaviors of human beings, thereby enhancing their utility and applicability as intelligent assistants.
Authors: Guangyu Chen, Yu Wu, Shujie Liu, Tao Liu, Xiaoyong Du, Furu Wei
Recent breakthroughs in zero-shot voice synthesis have enabled imitating a speaker's voice using just a few seconds of recording while maintaining a high level of realism. Alongside its potential benefits, this powerful technology introduces notable risks, including voice fraud and speaker impersonation. Unlike the conventional approach of solely relying on passive methods for detecting synthetic data, watermarking presents a proactive and robust defence mechanism against these looming risks. This paper introduces an innovative audio watermarking framework that encodes up to 32 bits of watermark within a mere 1-second audio snippet. The watermark is imperceptible to human senses and exhibits strong resilience against various attacks. It can serve as an effective identifier for synthesized voices and holds potential for broader applications in audio copyright protection. Moreover, this framework boasts high flexibility, allowing for the combination of multiple watermark segments to achieve heightened robustness and expanded capacity. Utilizing 10 to 20-second audio as the host, our approach demonstrates an average Bit Error Rate (BER) of 0.48\% across ten common attacks, a remarkable reduction of over 2800\% in BER compared to the state-of-the-art watermarking tool. See https://aka.ms/wavmark for demos of our work.
Authors: Chi Han, Qifan Wang, Wenhan Xiong, Yu Chen, Heng Ji, Sinong Wang
In recent years, there have been remarkable advancements in the performance of Transformer-based Large Language Models (LLMs) across various domains. As these LLMs are deployed for increasingly complex domains, they often face the need to follow longer user prompts or generate longer texts. In these situations, the $\textit{length generalization failure}$ of LLMs on long sequences becomes more prominent. Most pre-training schemes truncate training sequences to a fixed length. LLMs often struggle to generate fluent and coherent texts after longer contexts, even with relative positional encoding specifically designed to cope with this problem. Common solutions such as finetuning on longer corpora often involve daunting hardware and time costs and require careful training process design. To more efficiently extrapolate existing LLMs' generation quality to longer texts, we theoretically and empirically investigate the main out-of-distribution (OOD) factors contributing to this problem. Inspired by this diagnosis, we propose a simple yet effective solution for on-the-fly length generalization, LM-Infinite. It involves only a $\mathbf{\Lambda}$-shaped attention mask (to avoid excessive attended tokens) and a distance limit (to avoid unseen distances) while requiring no parameter updates or learning. We find it applicable to a variety of LLMs using relative-position encoding methods. LM-Infinite is computationally efficient with $O(n)$ time and space, and demonstrates consistent text generation fluency and quality to as long as 128k tokens on ArXiv and OpenWebText2 datasets, with 2.72x decoding speedup. We will make the codes publicly available following publication.
Authors: Luca Foppiano, Tomoya Mato, Kensei Terashima, Pedro Ortiz Suarez, Taku Tou, Chikako Sakai, Wei-Sheng Wang, Toshiyuki Amagasa, Yoshihiko Takano, Masashi Ishii
We propose a semi-automatic staging area for efficiently building an accurate database of experimental physical properties of superconductors from literature, called SuperCon2, to enrich the existing manually-built superconductor database SuperCon. Here we report our curation interface (SuperCon2 Interface) and a workflow managing the state transitions of each examined record, to validate the dataset of superconductors from PDF documents collected using Grobid-superconductors in a previous work. This curation workflow allows both automatic and manual operations, the former contains ``anomaly detection'' that scans new data identifying outliers, and a ``training data collector'' mechanism that collects training data examples based on manual corrections. Such training data collection policy is effective in improving the machine-learning models with a reduced number of examples. For manual operations, the interface (SuperCon2 interface) is developed to increase efficiency during manual correction by providing a smart interface and an enhanced PDF document viewer. We show that our interface significantly improves the curation quality by boosting precision and recall as compared with the traditional ``manual correction''. Our semi-automatic approach would provide a solution for achieving a reliable database with text-data mining of scientific documents.
Authors: Nadezhda Chirkova, Sheng Liang, Vassilina Nikoulina
Zero-shot cross-lingual generation assumes finetuning the multilingual pretrained language model (mPLM) on a generation task in one language and then using it to make predictions for this task in other languages. Previous works notice a frequent problem of generation in a wrong language and propose approaches to address it, usually using mT5 as a backbone model. In this work, we test alternative mPLMs, such as mBART and NLLB-200, and compare various approaches proposed in the literature in a unified setting. We first underline the importance of tuning learning rate used for finetuning, which helps to substantially alleviate the problem of generation in the wrong language. Then, we show that with careful learning rate tuning, the simple full finetuning of the model acts as a very strong baseline; other competitive approaches include parameter-efficient tuning with adapters and training on several source languages. Finally, we find that mBART performs similarly to mT5 of the same size, and NLLB-200 can be competitive in some cases.
Authors: Christina Chance, Da Yin, Dakuo Wang, Kai-Wei Chang
Recent studies show that traditional fairytales are rife with harmful gender biases. To help mitigate these gender biases in fairytales, this work aims to assess learned biases of language models by evaluating their robustness against gender perturbations. Specifically, we focus on Question Answering (QA) tasks in fairytales. Using counterfactual data augmentation to the FairytaleQA dataset, we evaluate model robustness against swapped gender character information, and then mitigate learned biases by introducing counterfactual gender stereotypes during training time. We additionally introduce a novel approach that utilizes the massive vocabulary of language models to support text genres beyond fairytales. Our experimental results suggest that models are sensitive to gender perturbations, with significant performance drops compared to the original testing set. However, when first fine-tuned on a counterfactual training dataset, models are less sensitive to the later introduced anti-gender stereotyped text.
Authors: Yaxin Fan, Feng Jiang, Benyou Wang, Peifeng Li, Haizhou Li
Foundation Models (FMs) have the potential to revolutionize the way users self-diagnose through search engines by offering direct and efficient suggestions. Recent studies primarily focused on the quality of FMs evaluated by GPT-4 or their ability to pass medical exams, no studies have quantified the extent of self-diagnostic atomic knowledge stored in FMs' memory, which is the basis of foundation models to provide factual and reliable suggestions. In this paper, we first constructed a benchmark of Self-diagnostic Atomic Knowledge (SdAK), including the most common types of atomic knowledge involved in self-diagnostic queries, with 17 atomic types and a total of 14, 048 pieces of atomic knowledge. Then, we evaluated both generic and open-source Chinese medical FMs on the benchmark. The experimental results showcase that generic FMs perform better than medical FMs in terms of self-diagnostic atomic knowledge. Error analysis revealed that both generic and medical FMs are sycophantic, e.g., always catering to users' claims when it comes to unknown knowledge. We further explored different types of data commonly adopted for fine-tuning medical FMs, i.e., real-world, semi-distilled, and distilled data, and found that distilled data can benefit FMs most. The code and data are available at \url{https://github.com/FreedomIntelligence/SDAK}.
Authors: Duc-Vu Nguyen, Quoc-Nam Nguyen
In this paper, we evaluate the ability of large language models (LLMs) to perform multiple choice symbol binding (MCSB) for multiple choice question answering (MCQA) tasks in zero-shot, one-shot, and few-shot settings. We focus on Vietnamese, with fewer challenging MCQA datasets than in English. The two existing datasets, ViMMRC 1.0 and ViMMRC 2.0, focus on literature. Recent research in Vietnamese natural language processing (NLP) has focused on the Vietnamese National High School Graduation Examination (VNHSGE) from 2019 to 2023 to evaluate ChatGPT. However, these studies have mainly focused on how ChatGPT solves the VNHSGE step by step. We aim to create a novel and high-quality dataset by providing structured guidelines for typing LaTeX formulas for mathematics, physics, chemistry, and biology. This dataset can be used to evaluate the MCSB ability of LLMs and smaller language models (LMs) because it is typed in a strict LaTeX style. We focus on predicting the character (A, B, C, or D) that is the most likely answer to a question, given the context of the question. Our evaluation of six well-known LLMs, namely BLOOMZ-7.1B-MT, LLaMA-2-7B, LLaMA-2-70B, GPT-3, GPT-3.5, and GPT-4.0, on the ViMMRC 1.0 and ViMMRC 2.0 benchmarks and our proposed dataset shows promising results on the MCSB ability of LLMs for Vietnamese. The dataset is available for research purposes only.
Authors: Devleena Das, Vivek Khetan
Recent advances have led to the availability of many pre-trained language models (PLMs); however, a question that remains is how much data is truly needed to fine-tune PLMs for downstream tasks? In this work, we introduce DEFT, a data-efficient fine-tuning framework that leverages unsupervised core-set selection to minimize the amount of data needed to fine-tune PLMs for downstream tasks. We demonstrate the efficacy of our DEFT framework in the context of text-editing LMs, and compare to the state-of-the art text-editing model, CoEDIT. Our quantitative and qualitative results demonstrate that DEFT models are just as accurate as CoEDIT while being finetuned on ~70% less data.
Authors: Joanne Boisson, Luis Espinosa-Anke, Jose Camacho-Collados
Metaphor identification aims at understanding whether a given expression is used figuratively in context. However, in this paper we show how existing metaphor identification datasets can be gamed by fully ignoring the potential metaphorical expression or the context in which it occurs. We test this hypothesis in a variety of datasets and settings, and show that metaphor identification systems based on language models without complete information can be competitive with those using the full context. This is due to the construction procedures to build such datasets, which introduce unwanted biases for positive and negative classes. Finally, we test the same hypothesis on datasets that are carefully sampled from natural corpora and where this bias is not present, making these datasets more challenging and reliable.
Authors: Xin Yuan, Jie Guo, Weidong Qiu, Zheng Huang, Shujun Li
Mis- and disinformation online have become a major societal problem as major sources of online harms of different kinds. One common form of mis- and disinformation is out-of-context (OOC) information, where different pieces of information are falsely associated, e.g., a real image combined with a false textual caption or a misleading textual description. Although some past studies have attempted to defend against OOC mis- and disinformation through external evidence, they tend to disregard the role of different pieces of evidence with different stances. Motivated by the intuition that the stance of evidence represents a bias towards different detection results, we propose a stance extraction network (SEN) that can extract the stances of different pieces of multi-modal evidence in a unified framework. Moreover, we introduce a support-refutation score calculated based on the co-occurrence relations of named entities into the textual SEN. Extensive experiments on a public large-scale dataset demonstrated that our proposed method outperformed the state-of-the-art baselines, with the best model achieving a performance gain of 3.2% in accuracy.
Authors: Olivia Huang, Eve Fleisig, Dan Klein
Current practices regarding data collection for natural language processing on Amazon Mechanical Turk (MTurk) often rely on a combination of studies on data quality and heuristics shared among NLP researchers. However, without considering the perspectives of MTurk workers, these approaches are susceptible to issues regarding workers' rights and poor response quality. We conducted a critical literature review and a survey of MTurk workers aimed at addressing open questions regarding best practices for fair payment, worker privacy, data quality, and considering worker incentives. We found that worker preferences are often at odds with received wisdom among NLP researchers. Surveyed workers preferred reliable, reasonable payments over uncertain, very high payments; reported frequently lying on demographic questions; and expressed frustration at having work rejected with no explanation. We also found that workers view some quality control methods, such as requiring minimum response times or Master's qualifications, as biased and largely ineffective. Based on the survey results, we provide recommendations on how future NLP studies may better account for MTurk workers' experiences in order to respect workers' rights and improve data quality.
Authors: Shaoguang Mao, Yuzhe Cai, Yan Xia, Wenshan Wu, Xun Wang, Fengyi Wang, Tao Ge, Furu Wei
This paper introduces Alympics, a platform that leverages Large Language Model (LLM) agents to facilitate investigations in game theory. By employing LLMs and autonomous agents to simulate human behavior and enable multi-agent collaborations, we can construct realistic and dynamic models of human interactions for game theory hypothesis formulating and testing. To demonstrate this, we present and implement a survival game involving unequal competition for limited resources. Through manipulation of resource availability and agent personalities, we observe how different agents engage in the competition and adapt their strategies. The use of LLM agents in game theory research offers significant advantages, including simulating realistic behavior, providing a controlled, scalable, and reproducible environment. Our work highlights the potential of LLM agents in enhancing the understanding of strategic decision-making within complex socioeconomic contexts. All codes are available at https://github.com/microsoft/Alympics
Authors: Chancharik Mitra, Abrar Anwar, Rodolfo Corona, Dan Klein, Trevor Darrell, Jesse Thomason
In this work, we consider the task of resolving object referents when given a comparative language description. We present a Multi-view Approach to Grounding in Context (MAGiC) that leverages transformers to pragmatically reason over both objects given multiple image views and a language description. In contrast to past efforts that attempt to connect vision and language for this task without fully considering the resulting referential context, MAGiC makes use of the comparative information by jointly reasoning over multiple views of both object referent candidates and the referring language expression. We present an analysis demonstrating that comparative reasoning contributes to SOTA performance on the SNARE object reference task.
Authors: Shuaijie She, Shujian Huang, Xingyun Wang, Yanke Zhou, Jiajun Chen
LLMs may interact with users in the form of dialogue and generate responses following their instructions, which naturally require dialogue comprehension abilities. However, dialogue comprehension is a general language ability which is hard to be evaluated directly. In this work, we propose to perform the evaluation with the help of the dialogue summarization task. Beside evaluating and analyzing the dialogue summarization performance (DIAC-Sum) of different LLMs, we also derive factual questions from the generated summaries and use them as a more flexible measurement of dialogue comprehension (DIAC-FactQA). Our evaluation shows that, on average, 27% of the summaries generated by LLMs contain factual inconsistency. Even ChatGPT, the strongest model evaluated, has such errors in 16% of its summaries. For answering the factual questions, which is more challenging, the average error rate of all evaluated LLMs is 37.2%. Both results indicate serious deficiencies. Detailed analysis shows that the understanding of subject/object of the conversation is still the most challenging problem for LLMs. Furthermore, to stimulate and enhance the dialogue comprehension ability of LLMs, we propose a fine-tuning paradigm with auto-constructed multi-task data. The experimental results demonstrate that our method achieved an error rate improvement of 10.9% on DIAC-FactQA.
Authors: Ang Lv, Kaiyi Zhang, Shufang Xie, Quan Tu, Yuhan Chen, Ji-Rong Wen, Rui Yan
Recent studies have highlighted a phenomenon in large language models (LLMs) known as "the reversal curse," in which the order of knowledge entities in the training data biases the models' comprehension. For example, if a model is trained on sentences where entity A consistently appears before entity B, it can respond to queries about A by providing B as the answer. However, it may encounter confusion when presented with questions concerning B. We contend that the reversal curse is partially a result of specific model training objectives, particularly evident in the prevalent use of the next-token prediction within most causal language models. For the next-token prediction, models solely focus on a token's preceding context, resulting in a restricted comprehension of the input. In contrast, we illustrate that the GLM, trained using the autoregressive blank infilling objective where tokens to be predicted have access to the entire context, exhibits better resilience against the reversal curse. We propose a novel training method, BIdirectional Casual language modeling Optimization (BICO), designed to mitigate the reversal curse when fine-tuning pretrained causal language models on new data. BICO modifies the causal attention mechanism to function bidirectionally and employs a mask denoising optimization. In the task designed to assess the reversal curse, our approach improves Llama's accuracy from the original 0% to around 70%. We hope that more attention can be focused on exploring and addressing these inherent weaknesses of the current LLMs, in order to achieve a higher level of intelligence.
Authors: Ken E. Friedl, Abbas Goher Khan, Soumya Ranjan Sahoo, Md Rashad Al Hasan Rony, Jana Germies, Christian Süß
The assessment of advanced generative large language models (LLMs) poses a significant challenge, given their heightened complexity in recent developments. Furthermore, evaluating the performance of LLM-based applications in various industries, as indicated by Key Performance Indicators (KPIs), is a complex undertaking. This task necessitates a profound understanding of industry use cases and the anticipated system behavior. Within the context of the automotive industry, existing evaluation metrics prove inadequate for assessing in-car conversational question answering (ConvQA) systems. The unique demands of these systems, where answers may relate to driver or car safety and are confined within the car domain, highlight the limitations of current metrics. To address these challenges, this paper introduces a set of KPIs tailored for evaluating the performance of in-car ConvQA systems, along with datasets specifically designed for these KPIs. A preliminary and comprehensive empirical evaluation substantiates the efficacy of our proposed approach. Furthermore, we investigate the impact of employing varied personas in prompts and found that it enhances the model's capacity to simulate diverse viewpoints in assessments, mirroring how individuals with different backgrounds perceive a topic.
Authors: C. Daniel Freeman, Laura Culp, Aaron Parisi, Maxwell L Bileschi, Gamaleldin F Elsayed, Alex Rizkowsky, Isabelle Simpson, Alex Alemi, Azade Nova, Ben Adlam, Bernd Bohnet, Gaurav Mishra, Hanie Sedghi, Igor Mordatch, Izzeddin Gur, Jaehoon Lee, JD Co-Reyes, Jeffrey Pennington, Kelvin Xu, Kevin Swersky, Kshiteej Mahajan, Lechao Xiao, Rosanne Liu, Simon Kornblith, Noah Constant, Peter J. Liu, Roman Novak, Yundi Qian, Noah Fiedel, Jascha Sohl-Dickstein
We introduce and study the problem of adversarial arithmetic, which provides a simple yet challenging testbed for language model alignment. This problem is comprised of arithmetic questions posed in natural language, with an arbitrary adversarial string inserted before the question is complete. Even in the simple setting of 1-digit addition problems, it is easy to find adversarial prompts that make all tested models (including PaLM2, GPT4, Claude2) misbehave, and even to steer models to a particular wrong answer. We additionally provide a simple algorithm for finding successful attacks by querying those same models, which we name "prompt inversion rejection sampling" (PIRS). We finally show that models can be partially hardened against these attacks via reinforcement learning and via agentic constitutional loops. However, we were not able to make a language model fully robust against adversarial arithmetic attacks.
Authors: Hanseok Oh, Haebin Shin, Miyoung Ko, Hyunji Lee, Minjoon Seo
We introduce a new problem KTRL+F, a knowledge-augmented in-document search task that necessitates real-time identification of all semantic targets within a document with the awareness of external sources through a single natural query. This task addresses following unique challenges for in-document search: 1) utilizing knowledge outside the document for extended use of additional information about targets to bridge the semantic gap between the query and the targets, and 2) balancing between real-time applicability with the performance. We analyze various baselines in KTRL+F and find there are limitations of existing models, such as hallucinations, low latency, or difficulties in leveraging external knowledge. Therefore we propose a Knowledge-Augmented Phrase Retrieval model that shows a promising balance between speed and performance by simply augmenting external knowledge embedding in phrase embedding. Additionally, we conduct a user study to verify whether solving KTRL+F can enhance search experience of users. It demonstrates that even with our simple model users can reduce the time for searching with less queries and reduced extra visits to other sources for collecting evidence. We encourage the research community to work on KTRL+F to enhance more efficient in-document information access.
Authors: Yifu Qiu, Zheng Zhao, Yftah Ziser, Anna Korhonen, Edoardo M. Ponti, Shay B. Cohen
Are Large language models (LLMs) temporally grounded? Since LLMs cannot perceive and interact with the environment, it is impossible to answer this question directly. Instead, we provide LLMs with textual narratives and probe them with respect to their common-sense knowledge of the structure and duration of events, their ability to order events along a timeline, and self-consistency within their temporal model (e.g., temporal relations such as after and before are mutually exclusive for any pair of events). We evaluate state-of-the-art LLMs (such as LLaMA 2 and GPT-4) on three tasks reflecting these abilities. Generally, we find that LLMs lag significantly behind both human performance as well as small-scale, specialised LMs. In-context learning, instruction tuning, and chain-of-thought prompting reduce this gap only to a limited degree. Crucially, LLMs struggle the most with self-consistency, displaying incoherent behaviour in at least 27.23% of their predictions. Contrary to expectations, we also find that scaling the model size does not guarantee positive gains in performance. To explain these results, we study the sources from which LLMs may gather temporal information: we find that sentence ordering in unlabelled texts, available during pre-training, is only weakly correlated with event ordering. Moreover, public instruction tuning mixtures contain few temporal tasks. Hence, we conclude that current LLMs lack a consistent temporal model of textual narratives. Code, datasets, and LLM outputs are available at https://github.com/yfqiu-nlp/temporal-llms.
Authors: Lin Xu, Zhiyuan Hu, Daquan Zhou, Hongyu Ren, Zhen Dong, Kurt Keutzer, See Kiong Ng, Jiashi Feng
Large Language Models (LLMs) have marked a significant advancement in the field of natural language processing, demonstrating exceptional capabilities in reasoning, tool usage, and memory. As their applications extend into multi-agent environments, a need has arisen for a comprehensive evaluation framework that captures their abilities in reasoning, planning, collaboration, and more. This work introduces a novel benchmarking framework specifically tailored to assess LLMs within multi-agent settings, providing quantitative metrics to evaluate their judgment, reasoning, deception, self-awareness, cooperation, coordination, and rationality. We utilize games such as Chameleon and Undercover, alongside game theory scenarios like Cost Sharing, Multi-player Prisoner's Dilemma, and Public Good, to create diverse testing environments. Our framework is fortified with the Probabilistic Graphical Modeling (PGM) method, enhancing the LLMs' capabilities in navigating complex social and cognitive dimensions. The benchmark evaluates seven multi-agent systems powered by different LLMs, quantitatively highlighting a significant capability gap over threefold between the strongest, GPT-4, and the weakest, Llama-2-70B. It also confirms that our PGM enhancement boosts the inherent abilities of all selected models by 50% on average. Our codes are released here https://github.com/cathyxl/MAgIC.
Authors: Yuxia Wang, Revanth Gangi Reddy, Zain Muhammad Mujahid, Arnav Arora, Aleksandr Rubashevskii, Jiahui Geng, Osama Mohammed Afzal, Liangming Pan, Nadav Borenstein, Aditya Pillai, Isabelle Augenstein, Iryna Gurevych, Preslav Nakov
The increased use of large language models (LLMs) across a variety of real-world applications calls for mechanisms to verify the factual accuracy of their outputs. In this work, we present a holistic end-to-end solution for annotating the factuality of LLM-generated responses, which encompasses a multi-stage annotation scheme designed to yield detailed labels concerning the verifiability and factual inconsistencies found in LLM outputs. We design and build an annotation tool to speed up the labelling procedure and ease the workload of raters. It allows flexible incorporation of automatic results in any stage, e.g. automatically-retrieved evidence. We further construct an open-domain document-level factuality benchmark in three-level granularity: claim, sentence and document. Preliminary experiments show that FacTool, FactScore and Perplexity.ai are struggling to identify false claims with the best F1=0.53. Annotation tool, benchmark and code are available at https://github.com/yuxiaw/Factcheck-GPT.
Authors: Zhaowei Wang, Haochen Shi, Weiqi Wang, Tianqing Fang, Hongming Zhang, Sehyun Choi, Xin Liu, Yangqiu Song
Cognitive research indicates that abstraction ability is essential in human intelligence, which remains under-explored in language models. In this paper, we present AbsPyramid, a unified entailment graph of 221K textual descriptions of abstraction knowledge. While existing resources only touch nouns or verbs within simplified events or specific domains, AbsPyramid collects abstract knowledge for three components of diverse events to comprehensively evaluate the abstraction ability of language models in the open domain. Experimental results demonstrate that current LLMs face challenges comprehending abstraction knowledge in zero-shot and few-shot settings. By training on our rich abstraction knowledge, we find LLMs can acquire basic abstraction abilities and generalize to unseen events. In the meantime, we empirically show that our benchmark is comprehensive to enhance LLMs across two previous abstraction tasks.