Authors: Kamal Acharya, Alvaro Velasquez, Houbing Herbert Song
Abstract: This survey paper delves into the emerging and critical area of symbolic knowledge distillation in Large Language Models (LLMs). As LLMs like Generative Pre-trained Transformer-3 (GPT-3) and Bidirectional Encoder Representations from Transformers (BERT) continue to expand in scale and complexity, the challenge of effectively harnessing their extensive knowledge becomes paramount. This survey concentrates on the process of distilling the intricate, often implicit knowledge contained within these models into a more symbolic, explicit form. This transformation is crucial for enhancing the interpretability, efficiency, and applicability of LLMs. We categorize the existing research based on methodologies and applications, focusing on how symbolic knowledge distillation can be used to improve the transparency and functionality of smaller, more efficient Artificial Intelligence (AI) models. The survey discusses the core challenges, including maintaining the depth of knowledge in a comprehensible format, and explores the various approaches and techniques that have been developed in this field. We identify gaps in current research and potential opportunities for future advancements. This survey aims to provide a comprehensive overview of symbolic knowledge distillation in LLMs, spotlighting its significance in the progression towards more accessible and efficient AI systems.
Authors: Weijia Zhang, Jia-Hong Huang, Svitlana Vakulenko, Yumo Xu, Thilina Rajapakse, Evangelos Kanoulas
Abstract: Query-focused summarization (QFS) is a fundamental task in natural language processing with broad applications, including search engines and report generation. However, traditional approaches assume the availability of relevant documents, which may not always hold in practical scenarios, especially in highly specialized topics. To address this limitation, we propose a novel knowledge-intensive approach that reframes QFS as a knowledge-intensive task setup. This approach comprises two main components: a retrieval module and a summarization controller. The retrieval module efficiently retrieves potentially relevant documents from a large-scale knowledge corpus based on the given textual query, eliminating the dependence on pre-existing document sets. The summarization controller seamlessly integrates a powerful large language model (LLM)-based summarizer with a carefully tailored prompt, ensuring the generated summary is comprehensive and relevant to the query. To assess the effectiveness of our approach, we create a new dataset, along with human-annotated relevance labels, to facilitate comprehensive evaluation covering both retrieval and summarization performance. Extensive experiments demonstrate the superior performance of our approach, particularly its ability to generate accurate summaries without relying on the availability of relevant documents initially. This underscores our method's versatility and practical applicability across diverse query scenarios.
Authors: Inkit Padhi, Karthikeyan Natesan Ramamurthy, Prasanna Sattigeri, Manish Nagireddy, Pierre Dognin, Kush R. Varshney
Abstract: Aligning large language models (LLMs) to value systems has emerged as a significant area of research within the fields of AI and NLP. Currently, this alignment process relies on the availability of high-quality supervised and preference data, which can be both time-consuming and expensive to curate or annotate. In this paper, we introduce a systematic end-to-end methodology for aligning LLMs to the implicit and explicit values represented in unstructured text data. Our proposed approach leverages the use of scalable synthetic data generation techniques to effectively align the model to the values present in the unstructured data. Through two distinct use-cases, we demonstrate the efficiency of our methodology on the Mistral-7B-Instruct model. Our approach credibly aligns LLMs to the values embedded within documents, and shows improved performance against other approaches, as quantified through the use of automatic metrics and win rates.
Authors: Hammad Rizwan, Domenic Rosati, Ga Wu, Hassan Sajjad
Abstract: Weight-preserving model editing techniques heavily rely on the scoping mechanism that decides when to apply an edit to the base model. These scoping mechanisms utilize distance functions in the representation space to ascertain the scope of the edit. In this work, we show that distance-based scoping functions grapple with lexical biases leading to issues such as misfires with irrelevant prompts that share similar lexical characteristics. To address this problem, we introduce, Projector Editor Networks for Model Editing (PENME),is a model editing approach that employs a compact adapter with a projection network trained via a contrastive learning objective. We demonstrate the efficacy of PENME in achieving superior results while being compute efficient and flexible to adapt across model architectures.
Authors: Tyler A. Chang, Catherine Arnett, Zhuowen Tu, Benjamin K. Bergen
Abstract: For many low-resource languages, the only available language models are large multilingual models trained on many languages simultaneously. However, using FLORES perplexity as a metric, we find that these models perform worse than bigrams for many languages (e.g. 24% of languages in XGLM 4.5B; 43% in BLOOM 7.1B). To facilitate research that focuses on low-resource languages, we pre-train and release Goldfish, a suite of monolingual autoregressive Transformer language models up to 125M parameters for 350 languages. The Goldfish reach lower FLORES perplexities than BLOOM, XGLM, and MaLA-500 on 98 of 204 FLORES languages, despite each Goldfish model being over 10x smaller. However, the Goldfish significantly underperform larger multilingual models on reasoning benchmarks, suggesting that for low-resource languages, multilinguality primarily improves general reasoning abilities rather than basic text generation. We release models trained on 5MB (350 languages), 10MB (288 languages), 100MB (166 languages), and 1GB (83 languages) of text data where available. The Goldfish models are available as baselines, fine-tuning sources, or augmentations to existing models in low-resource NLP research, and they are further useful for crosslinguistic studies requiring maximally comparable models across languages.
Authors: Guanchen Li, Xiandong Zhao, Lian Liu, Zeping Li, Dong Li, Lu Tian, Jie He, Ashish Sirasao, Emad Barsoum
Abstract: Pre-trained language models (PLMs) are engineered to be robust in contextual understanding and exhibit outstanding performance in various natural language processing tasks. However, their considerable size incurs significant computational and storage costs. Modern pruning strategies employ one-shot techniques to compress PLMs without the need for retraining on task-specific or otherwise general data; however, these approaches often lead to an indispensable reduction in performance. In this paper, we propose SDS, a Sparse-Dense-Sparse pruning framework to enhance the performance of the pruned PLMs from a weight distribution optimization perspective. We outline the pruning process in three steps. Initially, we prune less critical connections in the model using conventional one-shot pruning methods. Next, we reconstruct a dense model featuring a pruning-friendly weight distribution by reactivating pruned connections with sparse regularization. Finally, we perform a second pruning round, yielding a superior pruned model compared to the initial pruning. Experimental results demonstrate that SDS outperforms the state-of-the-art pruning techniques SparseGPT and Wanda under an identical sparsity configuration. For instance, SDS reduces perplexity by 9.13 on Raw-Wikitext2 and improves accuracy by an average of 2.05% across multiple zero-shot benchmarks for OPT-125M with 2:4 sparsity.
Authors: Ameya Godbole, Nicholas Monath, Seungyeon Kim, Ankit Singh Rawat, Andrew McCallum, Manzil Zaheer
Abstract: In text generation, hallucinations refer to the generation of seemingly coherent text that contradicts established knowledge. One compelling hypothesis is that hallucinations occur when a language model is given a generation task outside its parametric knowledge (due to rarity, recency, domain, etc.). A common strategy to address this limitation is to infuse the language models with retrieval mechanisms, providing the model with relevant knowledge for the task. In this paper, we leverage the planning capabilities of instruction-tuned LLMs and analyze how planning can be used to guide retrieval to further reduce the frequency of hallucinations. We empirically evaluate several variations of our proposed approach on long-form text generation tasks. By improving the coverage of relevant facts, plan-guided retrieval and generation can produce more informative responses while providing a higher rate of attribution to source documents.
Authors: Yihang Wang, Xu Huang, Bowen Tian, Yixing Fan, Jiafeng Guo
Abstract: Generative LLM have achieved significant success in various industrial tasks and can effectively adapt to vertical domains and downstream tasks through ICL. However, with tasks becoming increasingly complex, the context length required by ICL is also getting longer, and two significant issues arise: (i) The excessively long context leads to high costs and inference delays. (ii) A substantial amount of task-irrelevant information introduced by long contexts exacerbates the "lost in the middle" problem. Recently, compressing prompts by removing tokens according to some metric obtained from some causal language models, such as llama-7b, has emerged as an effective approach to mitigate these issues. However, the metric used by prior method such as self-information or PPL do not fully align with the objective of distinuishing the most important tokens when conditioning on query. In this work, we introduce information bottleneck theory to carefully examine the properties required by the metric. Inspired by this, we use cross-attention in encoder-decoder architecture as a new metric. Our simple method leads to significantly better performance in smaller models with lower latency. We evaluate our method on four datasets: DROP, CoQA, SQuAD, and Quoref. The experimental results show that, while maintaining the same performance, our compression rate can improve by nearly 25% over previous SOTA. Remarkably, in experiments where 25% of the tokens are removed, our model's EM score for answers sometimes even exceeds that of the control group using uncompressed text as context.
Authors: Zhiyang Qi, Michimasa Inaba
Abstract: This study addresses the interaction challenges encountered by spoken dialogue systems (SDSs) when engaging with users who exhibit distinct conversational behaviors, particularly minors, in scenarios where data are scarce. We propose a novel data augmentation framework to enhance SDS performance for user groups with limited resources. Our approach leverages a large language model (LLM) to extract speaker styles and a pre-trained language model (PLM) to simulate dialogue act history. This method generates enriched and personalized dialogue data, facilitating improved interactions with unique user demographics. Extensive experiments validate the efficacy of our methodology, highlighting its potential to foster the development of more adaptive and inclusive dialogue systems.
Authors: Xucheng Wan, Naijun Zheng, Kai Liu, Huan Zhou
Abstract: Contextualized ASR models have been demonstrated to effectively improve the recognition accuracy of uncommon phrases when a predefined phrase list is available. However, these models often struggle with bilingual settings, which are prevalent in code-switching speech recognition. In this study, we make the initial attempt to address this challenge by introducing a Cross-lingual Contextual Biasing(XCB) module. Specifically, we augment a pre-trained ASR model for the dominant language by integrating an auxiliary language biasing module and a supplementary language-specific loss, aimed at enhancing the recognition of phrases in the secondary language. Experimental results conducted on our in-house code-switching dataset have validated the efficacy of our approach, demonstrating significant improvements in the recognition of biasing phrases in the secondary language, even without any additional inference overhead. Additionally, our proposed system exhibits both efficiency and generalization when is applied by the unseen ASRU-2019 test set.
Authors: Tuc Nguyen, James Michels, Hua Shen, Thai Le
Abstract: In Explainable AI (XAI), counterfactual explanations (CEs) are a well-studied method to communicate feature relevance through contrastive reasoning of "what if" to explain AI models' predictions. However, they only focus on important (i.e., relevant) features and largely disregard less important (i.e., irrelevant) ones. Such irrelevant features can be crucial in many applications, especially when users need to ensure that an AI model's decisions are not affected or biased against specific attributes such as gender, race, religion, or political affiliation. To address this gap, the concept of alterfactual explanations (AEs) has been proposed. AEs explore an alternative reality of "no matter what", where irrelevant features are substituted with alternative features (e.g., "republicans" -> "democrats") within the same attribute (e.g., "politics") while maintaining a similar prediction output. This serves to validate whether AI model predictions are influenced by the specified attributes. Despite the promise of AEs, there is a lack of computational approaches to systematically generate them, particularly in the text domain, where creating AEs for AI text classifiers presents unique challenges. This paper addresses this challenge by formulating AE generation as an optimization problem and introducing MoMatterXAI, a novel algorithm that generates AEs for text classification tasks. Our approach achieves high fidelity of up to 95% while preserving context similarity of over 90% across multiple models and datasets. A human study further validates the effectiveness of AEs in explaining AI text classifiers to end users. All codes will be publicly available.
Authors: Yucheng Ruan, Xiang Lan, Jingying Ma, Yizhi Dong, Kai He, Mengling Feng
Abstract: Tabular data, a prevalent data type across various domains, presents unique challenges due to its heterogeneous nature and complex structural relationships. Achieving high predictive performance and robustness in tabular data analysis holds significant promise for numerous applications. Influenced by recent advancements in natural language processing, particularly transformer architectures, new methods for tabular data modeling have emerged. Early techniques concentrated on pre-training transformers from scratch, often encountering scalability issues. Subsequently, methods leveraging pre-trained language models like BERT have been developed, which require less data and yield enhanced performance. The recent advent of large language models, such as GPT and LLaMA, has further revolutionized the field, facilitating more advanced and diverse applications with minimal fine-tuning. Despite the growing interest, a comprehensive survey of language modeling techniques for tabular data remains absent. This paper fills this gap by providing a systematic review of the development of language modeling for tabular data, encompassing: (1) a categorization of different tabular data structures and data types; (2) a review of key datasets used in model training and tasks used for evaluation; (3) a summary of modeling techniques including widely-adopted data processing methods, popular architectures, and training objectives; (4) the evolution from adapting traditional Pre-training/Pre-trained language models to the utilization of large language models; (5) an identification of persistent challenges and potential future research directions in language modeling for tabular data analysis. GitHub page associated with this survey is available at: https://github.com/lanxiang1017/Language-Modeling-on-Tabular-Data-Survey.git.
URLs: https://github.com/lanxiang1017/Language-Modeling-on-Tabular-Data-Survey.git.
Authors: Hemant Yadav, Sunayana Sitaram, Rajiv Ratn Shah
Abstract: Speech modeling methods learn one embedding for a fixed segment of speech, typically in between 10-25 ms. The information present in speech can be divided into two categories: "what is being said" (content) and "how it is expressed" (other) and these two are orthogonal in nature causing the optimization algorithm to find a sub-optimal solution if forced to optimize together. This leads to sub-optimal performance in one or all downstream tasks as shown by previous studies. Current self-supervised learning (SSL) methods such as HuBERT are very good at modeling the content information present in speech. Data augmentation improves the performance on tasks which require effective modeling of other information but this leads to a divided capacity of the model. In this work, we conduct a preliminary study to understand the importance of modeling other information using separate learnable parameters. We propose a modified version of HuBERT, termed Other HuBERT (O-HuBERT), to test our hypothesis. Our findings are twofold: first, the O-HuBERT method is able to utilize all layers to build complex features to encode other information; second, a robust data augmentation strategy is essential for learning the information required by tasks that depend on other information and to achieve state-of-the-art (SOTA) performance on the SUPERB benchmark with a similarly sized model (100 million parameters) and pre-training data (960 hours).
Authors: Junhao Chen, Bowen Wang, Zhouqiang jiang, Yuta Nakashima
Abstract: Large Language Models (LLMs) have demonstrated significant capabilities, particularly in the domain of question answering (QA). However, their effectiveness in QA is often undermined by the vagueness of user questions. To address this issue, we introduce single-round instance-level prompt optimization, referred to as question rewriter. By enhancing the intelligibility of human questions for black-box LLMs, our question rewriter improves the quality of generated answers. The rewriter is optimized using direct preference optimization based on feedback collected from automatic criteria for evaluating generated answers; therefore, its training does not require costly human annotations. The experiments across multiple black-box LLMs and long-form question answering (LFQA) datasets demonstrate the efficacy of our method. This paper provides a practical framework for training question rewriters and sets a precedent for future explorations in prompt optimization within LFQA tasks. Code is available at \url{https://github.com/3244we/Question-Rewriter}.
Authors: Eui Jun Hwang, Sukmin Cho, Junmyeong Lee, Jong C. Park
Abstract: Gloss-free Sign Language Translation (SLT) converts sign videos directly into spoken language sentences without relying on glosses. Recently, Large Language Models (LLMs) have shown remarkable translation performance in gloss-free methods by harnessing their powerful natural language generation capabilities. However, these methods often rely on domain-specific fine-tuning of visual encoders to achieve optimal results. By contrast, this paper emphasizes the importance of capturing the spatial configurations and motion dynamics inherent in sign language. With this in mind, we introduce Spatial and Motion-based Sign Language Translation (SpaMo), a novel LLM-based SLT framework. The core idea of SpaMo is simple yet effective. We first extract spatial and motion features using off-the-shelf visual encoders and then input these features into an LLM with a language prompt. Additionally, we employ a visual-text alignment process as a warm-up before the SLT supervision. Our experiments demonstrate that SpaMo achieves state-of-the-art performance on two popular datasets, PHOENIX14T and How2Sign.
Authors: Ritwik Mishra, Sreeram Vennam, Rajiv Ratn Shah, Ponnurangam Kumaraguru
Abstract: Most existing Question Answering Datasets (QuADs) primarily focus on factoid-based short-context Question Answering (QA) in high-resource languages. However, the scope of such datasets for low-resource languages remains limited, with only a few works centered on factoid-based QuADs and none on non-factoid QuADs. Therefore, this work presents MuNfQuAD, a multilingual QuAD with non-factoid questions. It utilizes interrogative sub-headings from BBC news articles as questions and the corresponding paragraphs as silver answers. The dataset comprises over 370K QA pairs across 38 languages, encompassing several low-resource languages, and stands as the largest multilingual QA dataset to date. Based on the manual annotations of 790 QA-pairs from MuNfQuAD (golden set), we observe that 98\% of questions can be answered using their corresponding silver answer. Our fine-tuned Answer Paragraph Selection (APS) model outperforms the baselines. The APS model attained an accuracy of 80\% and 72\%, as well as a macro F1 of 72\% and 66\%, on the MuNfQuAD testset and the golden set, respectively. Furthermore, the APS model effectively generalizes certain a language within the golden set, even after being fine-tuned on silver labels.
Authors: Yongxin Deng (Shanghai University of Engineering Science), Xihe Qiu (Shanghai University of Engineering Science), Xiaoyu Tan (INF Technology), Jing Pan (Monash University), Chen Jue (Shanghai University of Engineering Science), Zhijun Fang (Donghua University), Yinghui Xu (Fudan University), Wei Chu (INF Technology), Yuan Qi (Fudan University)
Abstract: Large language models (LLMs) are trained on extensive text corpora, which inevitably include biased information. Although techniques such as Affective Alignment can mitigate some negative impacts of these biases, existing prompt-based attack methods can still extract these biases from the model's weights. Moreover, these biases frequently appear subtly when LLMs are prompted to perform identical tasks across different demographic groups, thereby camouflaging their presence. To address this issue, we have formally defined the implicit bias problem and developed an innovative framework for bias removal based on Bayesian theory, Bayesian-Theory based Bias Removal (BTBR). BTBR employs likelihood ratio screening to pinpoint data entries within publicly accessible biased datasets that represent biases inadvertently incorporated during the LLM training phase. It then automatically constructs relevant knowledge triples and expunges bias information from LLMs using model editing techniques. Through extensive experimentation, we have confirmed the presence of the implicit bias problem in LLMs and demonstrated the effectiveness of our BTBR approach.
Authors: Ming Jiang, Tingting Huang, Biao Guo, Yao Lu, Feng Zhang
Abstract: In recent years, Large language models (LLMs) have garnered significant attention due to their superior performance in complex reasoning tasks. However, recent studies may diminish their reasoning capabilities markedly when problem descriptions contain irrelevant information, even with the use of advanced prompting techniques. To further investigate this issue, a dataset of primary school mathematics problems containing irrelevant information, named GSMIR, was constructed. Testing prominent LLMs and prompting techniques on this dataset revealed that while LLMs can identify irrelevant information, they do not effectively mitigate the interference it causes once identified. A novel automatic construction method, ATF, which enhances the ability of LLMs to identify and self-mitigate the influence of irrelevant information, is proposed to address this shortcoming. This method operates in two steps: first, analysis of irrelevant information, followed by its filtering. The ATF method, as demonstrated by experimental results, significantly improves the reasoning performance of LLMs and prompting techniques, even in the presence of irrelevant information on the GSMIR dataset.
Authors: Maxim Ifergan, Leshem Choshen, Roee Aharoni, Idan Szpektor, Omri Abend
Abstract: The veracity of a factoid is largely independent of the language it is written in. However, language models are inconsistent in their ability to answer the same factual question across languages. This raises questions about how LLMs represent a given fact across languages. We explore multilingual factual knowledge through two aspects: the model's ability to answer a query consistently across languages, and the ability to ''store'' answers in a shared representation for several languages. We propose a methodology to measure the extent of representation sharing across languages by repurposing knowledge editing methods. We examine LLMs with various multilingual configurations using a new multilingual dataset. We reveal that high consistency does not necessarily imply shared representation, particularly for languages with different scripts. Moreover, we find that script similarity is a dominant factor in representation sharing. Finally, we observe that if LLMs could fully share knowledge across languages, their accuracy in their best-performing language could benefit an increase of up to 150\% on average. These findings highlight the need for improved multilingual knowledge representation in LLMs and suggest a path for the development of more robust and consistent multilingual LLMs.
Authors: Shu Chen, Xinyan Guan, Yaojie Lu, Hongyu Lin, Xianpei Han, Le Sun
Abstract: Manually annotating instruction data for large language models is difficult, costly, and hard to scale. Meanwhile, current automatic annotation methods typically rely on distilling synthetic data from proprietary LLMs, which not only limits the upper bound of the quality of the instruction data but also raises potential copyright issues. In this paper, we propose REInstruct, a simple and scalable method to automatically build instruction data from an unlabeled corpus without heavy reliance on proprietary LLMs and human annotation. Specifically, REInstruct first selects a subset of unlabeled texts that potentially contain well-structured helpful and insightful content and then generates instructions for these texts. To generate accurate and relevant responses for effective and robust training, REInstruct further proposes a rewriting-based approach to improve the quality of the generated instruction data. By training Llama-7b on a combination of 3k seed data and 32k synthetic data from REInstruct, fine-tuned model achieves a 65.41\% win rate on AlpacaEval leaderboard against text-davinci-003, outperforming other open-source, non-distilled instruction data construction methods. The code is publicly available at \url{https://github.com/cs32963/REInstruct}.
Authors: Tianyi Xu, Kaixun Huang, Pengcheng Guo, Yu Zhou, Longtao Huang, Hui Xue, Lei Xie
Abstract: Pre-trained multilingual speech foundation models, like Whisper, have shown impressive performance across different languages. However, adapting these models to new or specific languages is computationally extensive and faces catastrophic forgetting problems. Addressing these issues, our study investigates strategies to enhance the model on new languages in the absence of original training data, while also preserving the established performance on the original languages. Specifically, we first compare various LoRA-based methods to find out their vulnerability to forgetting. To mitigate this issue, we propose to leverage the LoRA parameters from the original model for approximate orthogonal gradient descent on the new samples. Additionally, we also introduce a learnable rank coefficient to allocate trainable parameters for more efficient training. Our experiments with a Chinese Whisper model (for Uyghur and Tibetan) yield better results with a more compact parameter set.
Authors: An Wang, Xingwu Sun, Ruobing Xie, Shuaipeng Li, Jiaqi Zhu, Zhen Yang, Pinxue Zhao, J. N. Han, Zhanhui Kang, Di Wang, Naoaki Okazaki, Cheng-zhong Xu
Abstract: Mixture of Experts (MoE) offers remarkable performance and computational efficiency by selectively activating subsets of model parameters. Traditionally, MoE models use homogeneous experts, each with identical capacity. However, varying complexity in input data necessitates experts with diverse capabilities, while homogeneous MoE hinders effective expert specialization and efficient parameter utilization. In this study, we propose a novel Heterogeneous Mixture of Experts (HMoE), where experts differ in size and thus possess diverse capacities. This heterogeneity allows for more specialized experts to handle varying token complexities more effectively. To address the imbalance in expert activation, we propose a novel training objective that encourages the frequent activation of smaller experts, enhancing computational efficiency and parameter utilization. Extensive experiments demonstrate that HMoE achieves lower loss with fewer activated parameters and outperforms conventional homogeneous MoE models on various pre-training evaluation benchmarks. Codes will be released upon acceptance.
Authors: Hongbang Yuan, Zhuoran Jin, Pengfei Cao, Yubo Chen, Kang Liu, Jun Zhao
Abstract: LLM have achieved success in many fields but still troubled by problematic content in the training corpora. LLM unlearning aims at reducing their influence and avoid undesirable behaviours. However, existing unlearning methods remain vulnerable to adversarial queries and the unlearned knowledge resurfaces after the manually designed attack queries. As part of a red-team effort to proactively assess the vulnerabilities of unlearned models, we design Dynamic Unlearning Attack (DUA), a dynamic and automated framework to attack these models and evaluate their robustness. It optimizes adversarial suffixes to reintroduce the unlearned knowledge in various scenarios. We find that unlearned knowledge can be recovered in $55.2\%$ of the questions, even without revealing the unlearned model's parameters. In response to this vulnerability, we propose Latent Adversarial Unlearning (LAU), a universal framework that effectively enhances the robustness of the unlearned process. It formulates the unlearning process as a min-max optimization problem and resolves it through two stages: an attack stage, where perturbation vectors are trained and added to the latent space of LLMs to recover the unlearned knowledge, and a defense stage, where previously trained perturbation vectors are used to enhance unlearned model's robustness. With our LAU framework, we obtain two robust unlearning methods, AdvGA and AdvNPO. We conduct extensive experiments across multiple unlearning benchmarks and various models, and demonstrate that they improve the unlearning effectiveness by over $53.5\%$, cause only less than a $11.6\%$ reduction in neighboring knowledge, and have almost no impact on the model's general capabilities.
Authors: Artem Vazhentsev, Ekaterina Fadeeva, Rui Xing, Alexander Panchenko, Preslav Nakov, Timothy Baldwin, Maxim Panov, Artem Shelmanov
Abstract: Uncertainty quantification (UQ) is a perspective approach to detecting Large Language Model (LLM) hallucinations and low quality output. In this work, we address one of the challenges of UQ in generation tasks that arises from the conditional dependency between the generation steps of an LLM. We propose to learn this dependency from data. We train a regression model, which target variable is the gap between the conditional and the unconditional generation confidence. During LLM inference, we use this learned conditional dependency model to modulate the uncertainty of the current generation step based on the uncertainty of the previous step. Our experimental evaluation on nine datasets and three LLMs shows that the proposed method is highly effective for uncertainty quantification, achieving substantial improvements over rivaling approaches.
Authors: Tej Deep Pala, Vernon Y. H. Toh, Rishabh Bhardwaj, Soujanya Poria
Abstract: In today's era, where large language models (LLMs) are integrated into numerous real-world applications, ensuring their safety and robustness is crucial for responsible AI usage. Automated red-teaming methods play a key role in this process by generating adversarial attacks to identify and mitigate potential vulnerabilities in these models. However, existing methods often struggle with slow performance, limited categorical diversity, and high resource demands. While Rainbow Teaming, a recent approach, addresses the diversity challenge by framing adversarial prompt generation as a quality-diversity search, it remains slow and requires a large fine-tuned mutator for optimal performance. To overcome these limitations, we propose Ferret, a novel approach that builds upon Rainbow Teaming by generating multiple adversarial prompt mutations per iteration and using a scoring function to rank and select the most effective adversarial prompt. We explore various scoring functions, including reward models, Llama Guard, and LLM-as-a-judge, to rank adversarial mutations based on their potential harm to improve the efficiency of the search for harmful mutations. Our results demonstrate that Ferret, utilizing a reward model as a scoring function, improves the overall attack success rate (ASR) to 95%, which is 46% higher than Rainbow Teaming. Additionally, Ferret reduces the time needed to achieve a 90% ASR by 15.2% compared to the baseline and generates adversarial prompts that are transferable i.e. effective on other LLMs of larger size. Our codes are available at https://github.com/declare-lab/ferret.
Authors: Jiyang Qiu, Xinbei Ma, Zhuosheng Zhang, Hai Zhao
Abstract: Large language models (LLMs) have demonstrated remarkable capabilities. Their powerful generative abilities enable flexible responses based on various queries or instructions. Emerging as widely adopted generalists for diverse tasks, LLMs are still vulnerable to backdoors. This paper proposes an editing-based generative backdoor, named MEGen, aiming to create a customized backdoor for NLP tasks with the least side effects. In our approach, we first leverage a language model to insert a trigger selected on fixed metrics into the input, then design a pipeline of model editing to directly embed a backdoor into an LLM. By adjusting a small set of local parameters with a mini-batch of samples, MEGen significantly enhances time efficiency and achieves high robustness. Experimental results indicate that our backdoor attack strategy achieves a high attack success rate on poison data while maintaining the model's performance on clean data. Notably, the backdoored model, when triggered, can freely output pre-set dangerous information while successfully completing downstream tasks. This suggests that future LLM applications could be guided to deliver certain dangerous information, thus altering the LLM's generative style. We believe this approach provides insights for future LLM applications and the execution of backdoor attacks on conversational AI systems.
Authors: Cem \"Uy\"uk, Danica Rov\'o, Shaghayegh Kolli, Rabia Varol, Georg Groh, Daryna Dementieva
Abstract: In the era dominated by information overload and its facilitation with Large Language Models (LLMs), the prevalence of misinformation poses a significant threat to public discourse and societal well-being. A critical concern at present involves the identification of machine-generated news. In this work, we take a significant step by introducing a benchmark dataset designed for neural news detection in four languages: English, Turkish, Hungarian, and Persian. The dataset incorporates outputs from multiple multilingual generators (in both, zero-shot and fine-tuned setups) such as BloomZ, LLaMa-2, Mistral, Mixtral, and GPT-4. Next, we experiment with a variety of classifiers, ranging from those based on linguistic features to advanced Transformer-based models and LLMs prompting. We present the detection results aiming to delve into the interpretablity and robustness of machine-generated texts detectors across all target languages.
Authors: Huy Quoc To, Ming Liu, Guangyan Huang
Abstract: Large language models (LLMs) have ushered in a new era for processing complex information in various fields, including science. The increasing amount of scientific literature allows these models to acquire and understand scientific knowledge effectively, thus improving their performance in a wide range of tasks. Due to the power of LLMs, they require extremely expensive computational resources, intense amounts of data, and training time. Therefore, in recent years, researchers have proposed various methodologies to make scientific LLMs more affordable. The most well-known approaches align in two directions. It can be either focusing on the size of the models or enhancing the quality of data. To date, a comprehensive review of these two families of methods has not yet been undertaken. In this paper, we (I) summarize the current advances in the emerging abilities of LLMs into more accessible AI solutions for science, and (II) investigate the challenges and opportunities of developing affordable solutions for scientific domains using LLMs.
Authors: Chenhan Yuan, Fei Huang, Ru Peng, Keming Lu, Bowen Yu, Chang Zhou, Jingren Zhou
Abstract: Transformer-based large language models (LLMs) exhibit limitations such as generating unsafe responses, unreliable reasoning, etc. Existing inference intervention approaches attempt to mitigate these issues by finetuning additional models to produce calibration signals (such as rewards) that guide the LLM's decoding process. However, this solution introduces substantial time and space overhead due to the separate models required. This work proposes Non-disruptive parameters insertion (Otter), inserting extra parameters into the transformer architecture to predict calibration signals along with the original LLM output. Otter offers state-of-the-art performance on multiple demanding tasks while saving up to 86.5\% extra space and 98.5\% extra time. Furthermore, Otter seamlessly integrates with existing inference engines, requiring only a one-line code change, and the original model response remains accessible after the parameter insertion. Our code is publicly available at \url{https://github.com/chenhan97/Otter}
Authors: Yuankai Zhang, Lingxiao Kong, Haozhao Wang, Ruixuan Li, Jun Wang, Yuhua Li, Wei Liu
Abstract: Rationalization models, which select a subset of input text as rationale-crucial for humans to understand and trust predictions-have recently emerged as a prominent research area in eXplainable Artificial Intelligence. However, most of previous studies mainly focus on improving the quality of the rationale, ignoring its robustness to malicious attack. Specifically, whether the rationalization models can still generate high-quality rationale under the adversarial attack remains unknown. To explore this, this paper proposes UAT2E, which aims to undermine the explainability of rationalization models without altering their predictions, thereby eliciting distrust in these models from human users. UAT2E employs the gradient-based search on triggers and then inserts them into the original input to conduct both the non-target and target attack. Experimental results on five datasets reveal the vulnerability of rationalization models in terms of explanation, where they tend to select more meaningless tokens under attacks. Based on this, we make a series of recommendations for improving rationalization models in terms of explanation.
Authors: Alex Gichamba, Tewodros Kederalah Idris, Brian Ebiyau, Eric Nyberg, Teruko Mitamura
Abstract: Domain-specific question answering remains challenging for language models, given the deep technical knowledge required to answer questions correctly. This difficulty is amplified for smaller language models that cannot encode as much information in their parameters as larger models. The "Specializing Large Language Models for Telecom Networks" challenge aimed to enhance the performance of two small language models, Phi-2 and Falcon-7B in telecommunication question answering. In this paper, we present our question answering systems for this challenge. Our solutions achieved leading marks of 81.9% accuracy for Phi-2 and 57.3% for Falcon-7B. We have publicly released our code and fine-tuned models.
Authors: Chengzhi Zhong, Fei Cheng, Qianying Liu, Junfeng Jiang, Zhen Wan, Chenhui Chu, Yugo Murawaki, Sadao Kurohashi
Abstract: In this study, we investigate whether non-English-centric LLMs, despite their strong performance, `think' in their respective dominant language: more precisely, `think' refers to how the representations of intermediate layers, when un-embedded into the vocabulary space, exhibit higher probabilities for certain dominant languages during generation. We term such languages as internal $\textbf{latent languages}$. We examine the latent language of three typical categories of models for Japanese processing: Llama2, an English-centric model; Swallow, an English-centric model with continued pre-training in Japanese; and LLM-jp, a model pre-trained on balanced English and Japanese corpora. Our empirical findings reveal that, unlike Llama2 which relies exclusively on English as the internal latent language, Japanese-specific Swallow and LLM-jp employ both Japanese and English, exhibiting dual internal latent languages. For any given target language, the model preferentially activates the latent language most closely related to it. In addition, we explore how intermediate layers respond to questions involving cultural conflicts between latent internal and target output languages. We further explore how the language identity shifts across layers while keeping consistent semantic meaning reflected in the intermediate layer representations. This study deepens the understanding of non-English-centric large language models, highlighting the intricate dynamics of language representation within their intermediate layers.
Authors: Rui Yang, Jiahao Zhu, Jianping Man, Li Fang, Yi Zhou
Abstract: Knowledge graph completion (KGC) aims to identify missing triples in a knowledge graph (KG). This is typically achieved through tasks such as link prediction and instance completion. However, these methods often focus on either static knowledge graphs (SKGs) or temporal knowledge graphs (TKGs), addressing only within-scope triples. This paper introduces a new generative completion framework called Generative Subgraph-based KGC (GS-KGC). GS-KGC employs a question-answering format to directly generate target entities, addressing the challenge of questions having multiple possible answers. We propose a strategy that extracts subgraphs centered on entities and relationships within the KG, from which negative samples and neighborhood information are separately obtained to address the one-to-many problem. Our method generates negative samples using known facts to facilitate the discovery of new information. Furthermore, we collect and refine neighborhood path data of known entities, providing contextual information to enhance reasoning in large language models (LLMs). Our experiments evaluated the proposed method on four SKGs and two TKGs, achieving state-of-the-art Hits@1 metrics on five datasets. Analysis of the results shows that GS-KGC can discover new triples within existing KGs and generate new facts beyond the closed KG, effectively bridging the gap between closed-world and open-world KGC.
Authors: Kathrin Se{\ss}ler, Yao Rong, Emek G\"ozl\"ukl\"u, Enkelejda Kasneci
Abstract: The use of Large Language Models (LLMs) in mathematical reasoning has become a cornerstone of related research, demonstrating the intelligence of these models and enabling potential practical applications through their advanced performance, such as in educational settings. Despite the variety of datasets and in-context learning algorithms designed to improve the ability of LLMs to automate mathematical problem solving, the lack of comprehensive benchmarking across different datasets makes it complicated to select an appropriate model for specific tasks. In this project, we present a benchmark that fairly compares seven state-of-the-art in-context learning algorithms for mathematical problem solving across five widely used mathematical datasets on four powerful foundation models. Furthermore, we explore the trade-off between efficiency and performance, highlighting the practical applications of LLMs for mathematical reasoning. Our results indicate that larger foundation models like GPT-4o and LLaMA 3-70B can solve mathematical reasoning independently from the concrete prompting strategy, while for smaller models the in-context learning approach significantly influences the performance. Moreover, the optimal prompt depends on the chosen foundation model. We open-source our benchmark code to support the integration of additional models in future research.
Authors: John Mendon\c{c}a, Isabel Trancoso, Alon Lavie
Abstract: Although human evaluation remains the gold standard for open-domain dialogue evaluation, the growing popularity of automated evaluation using Large Language Models (LLMs) has also extended to dialogue. However, most frameworks leverage benchmarks that assess older chatbots on aspects such as fluency and relevance, which are not reflective of the challenges associated with contemporary models. In fact, a qualitative analysis on Soda, a GPT-3.5 generated dialogue dataset, suggests that current chatbots may exhibit several recurring issues related to coherence and commonsense knowledge, but generally produce highly fluent and relevant responses. Noting the aforementioned limitations, this paper introduces Soda-Eval, an annotated dataset based on Soda that covers over 120K turn-level assessments across 10K dialogues, where the annotations were generated by GPT-4. Using Soda-Eval as a benchmark, we then study the performance of several open-access instruction-tuned LLMs, finding that dialogue evaluation remains challenging. Fine-tuning these models improves performance over few-shot inferences, both in terms of correlation and explanation.
Authors: Yeyong Yu, Rusheng Yu, Haojie Wei, Zhanqiu Zhang, Quan Qian
Abstract: The rapid advancement of large language models (LLMs) has revolutionized role-playing, enabling the development of general role-playing models. However, current role-playing training has two significant issues: (I) Using a predefined role profile to prompt dialogue training for specific scenarios usually leads to inconsistencies and even conflicts between the dialogue and the profile, resulting in training biases. (II) The model learns to imitate the role based solely on the profile, neglecting profile-dialogue alignment at the sentence level. In this work, we propose a simple yet effective framework called BEYOND DIALOGUE, designed to overcome these hurdles. This framework innovatively introduces "beyond dialogue" tasks to align dialogue with profile traits based on each specific scenario, thereby eliminating biases during training. Furthermore, by adopting an innovative prompting mechanism that generates reasoning outcomes for training, the framework allows the model to achieve fine-grained alignment between profile and dialogue at the sentence level. The aforementioned methods are fully automated and low-cost. Additionally, the integration of automated dialogue and objective evaluation methods forms a comprehensive framework, paving the way for general role-playing. Experimental results demonstrate that our model excels in adhering to and reflecting various dimensions of role profiles, outperforming most proprietary general and specialized role-playing baselines. All code and datasets are available at https://github.com/yuyouyu32/BeyondDialogue.
Authors: Viraat Aryabumi, Yixuan Su, Raymond Ma, Adrien Morisot, Ivan Zhang, Acyr Locatelli, Marzieh Fadaee, Ahmet \"Ust\"un, Sara Hooker
Abstract: Including code in the pre-training data mixture, even for models not specifically designed for code, has become a common practice in LLMs pre-training. While there has been anecdotal consensus among practitioners that code data plays a vital role in general LLMs' performance, there is only limited work analyzing the precise impact of code on non-code tasks. In this work, we systematically investigate the impact of code data on general performance. We ask "what is the impact of code data used in pre-training on a large variety of downstream tasks beyond code generation". We conduct extensive ablations and evaluate across a broad range of natural language reasoning tasks, world knowledge tasks, code benchmarks, and LLM-as-a-judge win-rates for models with sizes ranging from 470M to 2.8B parameters. Across settings, we find a consistent results that code is a critical building block for generalization far beyond coding tasks and improvements to code quality have an outsized impact across all tasks. In particular, compared to text-only pre-training, the addition of code results in up to relative increase of 8.2% in natural language (NL) reasoning, 4.2% in world knowledge, 6.6% improvement in generative win-rates, and a 12x boost in code performance respectively. Our work suggests investments in code quality and preserving code during pre-training have positive impacts.
Authors: Jiasheng Si, Yibo Zhao, Yingjie Zhu, Haiyang Zhu, Wenpeng Lu, Deyu Zhou
Abstract: With the growing complexity of fact verification tasks, the concern with "thoughtful" reasoning capabilities is increasing. However, recent fact verification benchmarks mainly focus on checking a narrow scope of semantic factoids within claims and lack an explicit logical reasoning process. In this paper, we introduce CheckWhy, a challenging dataset tailored to a novel causal fact verification task: checking the truthfulness of the causal relation within claims through rigorous reasoning steps. CheckWhy consists of over 19K "why" claim-evidence-argument structure triplets with supports, refutes, and not enough info labels. Each argument structure is composed of connected evidence, representing the reasoning process that begins with foundational evidence and progresses toward claim establishment. Through extensive experiments on state-of-the-art models, we validate the importance of incorporating the argument structure for causal fact verification. Moreover, the automated and human evaluation of argument structure generation reveals the difficulty in producing satisfying argument structure by fine-tuned models or Chain-of-Thought prompted LLMs, leaving considerable room for future improvements.
Authors: Kangjun Noh, Baekryun Seong, Hoyoon Byun, Sungjin Song, Kyungwoo Song
Abstract: Large Language Models (LLMs) have great success in natural language processing tasks such as response generation. However, their use in tabular data has been limited due to their inferior performance compared to traditional machine learning models (TMLs) such as XGBoost. We find that the pre-trained knowledge of LLMs enables them to interpret new variables that appear in a test without additional training, a capability central to the concept of Out-of-Variable (OOV). From the findings, we propose a Language-Based-Classifier (LBC), a classifier that maximizes the benefits of LLMs to outperform TMLs on OOV tasks. LBC employs three key methodological strategies: 1) Categorical changes to adjust data to better fit the model's understanding, 2) Advanced order and indicator to enhance data representation to the model, and 3) Using verbalizer to map logit scores to classes during inference to generate model predictions. These strategies, combined with the pre-trained knowledge of LBC, emphasize the model's ability to effectively handle OOV tasks. We empirically and theoretically validate the superiority of LBC. LBC is the first study to apply an LLM-based model to OOV tasks. The source code is at https://github.com/ASDASDanonymous/Language-Based-Classifier-forOOVtasks.
URLs: https://github.com/ASDASDanonymous/Language-Based-Classifier-forOOVtasks.
Authors: Yanzhao Qin, Tao Zhang, Tao Zhang, Yanjun Shen, Wenjing Luo, Haoze Sun, Yan Zhang, Yujing Qiao, Weipeng Chen, Zenan Zhou, Wentao Zhang, Bin Cui
Abstract: Large Language Models (LLMs) have become instrumental across various applications, with the customization of these models to specific scenarios becoming increasingly critical. System message, a fundamental component of LLMs, is consist of carefully crafted instructions that guide the behavior of model to meet intended goals. Despite the recognized potential of system messages to optimize AI-driven solutions, there is a notable absence of a comprehensive benchmark for evaluating how well different LLMs follow these system messages. To fill this gap, we introduce SysBench, a benchmark that systematically analyzes system message following ability in terms of three challenging aspects: constraint complexity, instruction misalignment and multi-turn stability. In order to enable effective evaluation, SysBench constructs multi-turn user conversations covering various interaction relationships, based on six common types of constraints from system messages in real-world scenarios. Our dataset contains 500 system messages from various domains, each paired with 5 turns of user conversations, which have been manually formulated and checked to guarantee high quality. SysBench provides extensive evaluation across various LLMs, measuring their ability to follow specified constraints given in system messages. The results highlight both the strengths and weaknesses of existing models, offering key insights and directions for future research. The open source library SysBench is available at https://github.com/PKU-Baichuan-MLSystemLab/SysBench.
Authors: Katerina Papantoniou, Yannis Tzitzikas
Abstract: English language is in the spotlight of the Natural Language Processing (NLP) community with other languages, like Greek, lagging behind in terms of offered methods, tools and resources. Due to the increasing interest in NLP, in this paper we try to condense research efforts for the automatic processing of Greek language covering the last three decades. In particular, we list and briefly discuss related works, resources and tools, categorized according to various processing layers and contexts. We are not restricted to the modern form of Greek language but also cover Ancient Greek and various Greek dialects. This survey can be useful for researchers and students interested in NLP tasks, Information Retrieval and Knowledge Management for the Greek language.
Authors: Hongfa Zi, Zhen Liu
Abstract: The characters used in modern countries are mainly divided into ideographic characters and phonetic characters, both of which have their advantages and disadvantages. Chinese is difficult to learn and easy to master, while English is easy to learn but has a large vocabulary. There is still no language that combines the advantages of both languages and has less memory capacity, can form words, and is easy to learn. Therefore, inventing new characters that can be combined and the popularization of deep knowledge, and reduce disputes through communication. Firstly, observe the advantages and disadvantages of Chinese and English, such as their vocabulary, information content, and ease of learning in deep scientific knowledge, and create a new writing system. Then, use comparative analysis to observe the total score of the new language. Through this article, it can be concluded that the new text combines the advantages of both pictographic and alphabetical writing: new characters that can be combined into words reduces the vocabulary that needs to be learned; Special prefixes allow beginners to quickly guess the approximate category and meaning of unseen words; New characters can enable humans to quickly learn more advanced knowledge.
Authors: Michael Reinisch, Jianfeng He, Chenxi Liao, Sauleh Ahmad Siddiqui, Bei Xiao
Abstract: New medical treatment development requires multiple phases of clinical trials. Despite the significant human and financial costs of bringing a drug to market, less than 20% of drugs in testing will make it from the first phase to final approval. Recent literature indicates that the design of the trial protocols significantly contributes to trial performance. We investigated Clinical Trial Outcome Prediction (CTOP) using trial design documents to predict phase transitions automatically. We propose CTP-LLM, the first Large Language Model (LLM) based model for CTOP. We also introduce the PhaseTransition (PT) Dataset; which labels trials based on their progression through the regulatory process and serves as a benchmark for CTOP evaluation. Our fine-tuned GPT-3.5-based model (CTP-LLM) predicts clinical trial phase transition by analyzing the trial's original protocol texts without requiring human-selected features. CTP-LLM achieves a 67% accuracy rate in predicting trial phase transitions across all phases and a 75% accuracy rate specifically in predicting the transition from Phase~III to final approval. Our experimental performance highlights the potential of LLM-powered applications in forecasting clinical trial outcomes and assessing trial design.
Authors: Waris Quamer, Ricardo Gutierrez-Osuna
Abstract: Current accent conversion (AC) systems do not disentangle the two main sources of non-native accent: segmental and prosodic characteristics. Being able to manipulate a non-native speaker's segmental and/or prosodic channels independently is critical to quantify how these two channels contribute to speech comprehensibility and social attitudes. We present an AC system that not only decouples voice quality from accent, but also disentangles the latter into its segmental and prosodic characteristics. The system is able to generate accent conversions that combine (1) the segmental characteristics from a source utterance, (2) the voice characteristics from a target utterance, and (3) the prosody of a reference utterance. We show that vector quantization of acoustic embeddings and removal of consecutive duplicated codewords allows the system to transfer prosody and improve voice similarity. We conduct perceptual listening tests to quantify the individual contributions of segmental features and prosody on the perceived comprehensibility of non-native speech. Our results indicate that, contrary to prior research in non-native speech, segmental features have a larger impact on comprehensibility than prosody. The proposed AC system may also be used to study how segmental and prosody cues affect social attitudes towards non-native speech.
Authors: Wen Cheng, Ke Sun, Xinyu Zhang, Wei Wang
Abstract: The rapid development of large language models (LLMs) has significantly advanced code completion capabilities, giving rise to a new generation of LLM-based Code Completion Tools (LCCTs). Unlike general-purpose LLMs, these tools possess unique workflows, integrating multiple information sources as input and prioritizing code suggestions over natural language interaction, which introduces distinct security challenges. Additionally, LCCTs often rely on proprietary code datasets for training, raising concerns about the potential exposure of sensitive data. This paper exploits these distinct characteristics of LCCTs to develop targeted attack methodologies on two critical security risks: jailbreaking and training data extraction attacks. Our experimental results expose significant vulnerabilities within LCCTs, including a 99.4% success rate in jailbreaking attacks on GitHub Copilot and a 46.3% success rate on Amazon Q. Furthermore, We successfully extracted sensitive user data from GitHub Copilot, including 54 real email addresses and 314 physical addresses associated with GitHub usernames. Our study also demonstrates that these code-based attack methods are effective against general-purpose LLMs, such as the GPT series, highlighting a broader security misalignment in the handling of code by modern LLMs. These findings underscore critical security challenges associated with LCCTs and suggest essential directions for strengthening their security frameworks. The example code and attack samples from our research are provided at https://github.com/Sensente/Security-Attacks-on-LCCTs.
URLs: https://github.com/Sensente/Security-Attacks-on-LCCTs.
Authors: Tanmana Sadhu, Ali Pesaranghader, Yanan Chen, Dong Hoon Yi
Abstract: Due to emergent capabilities, large language models (LLMs) have been utilized as language-based agents to perform a variety of tasks and make decisions with an increasing degree of autonomy. These autonomous agents can understand high-level instructions, interact with their environments, and execute complex tasks using a selection of tools available to them. As the capabilities of the agents expand, ensuring their safety and trustworthiness becomes more imperative. In this study, we introduce the Athena framework which leverages the concept of verbal contrastive learning where past safe and unsafe trajectories are used as in-context (contrastive) examples to guide the agent towards safety while fulfilling a given task. The framework also incorporates a critiquing mechanism to guide the agent to prevent risky actions at every step. Furthermore, due to the lack of existing benchmarks on the safety reasoning ability of LLM-based agents, we curate a set of 80 toolkits across 8 categories with 180 scenarios to provide a safety evaluation benchmark. Our experimental evaluation, with both closed- and open-source LLMs, indicates verbal contrastive learning and interaction-level critiquing improve the safety rate significantly.
Authors: Howe Tissue, Venus Wang, Lu Wang
Abstract: We find that the cross-entropy loss curves of neural language models empirically adhere to a scaling law with learning rate (LR) annealing over training steps ($s$): $$L(s) = L_0 + A\cdot S_1^{-\alpha} - C\cdot S_2$$ Where $S_1$ is forward area and $S_2$ is learning rate annealing area. This formulation takes into account two factors: (1) The forward scaling defined as typical scaling law, and (2) the additional loss drop brought by LR annealing. Therefore, this formulation can describe the full loss curve at each step, rather than the single loss point at the end of training. Applying the scaling law with LR annealing and fitting only one or two training curves, we can accurately predict the loss of language model training at any given step and across any learning rate scheduler (LRS). Furthermore, this equation accurately describes the dynamics during training process, and provides a theoretical verification and explanation for numerous experimental findings of previous studies, particularly those focusing on LR schedule and LR annealing. The resulting insights, also serve as a guide for researchers to select critical LRS in advance by prediction using our equation. Most significantly, since all the points in a full training curve follow the equation, we can achieve accurate loss prediction at any given step across any learning rate scheduler, while expending less than 1\% of the computational cost required by the chinchilla scaling law to fit language modeling loss. This approach extremely democratizes scaling law fitting and predicting in developing large language models.
Authors: Yuan Xin, Zheng Li, Ning Yu, Dingfan Chen, Mario Fritz, Michael Backes, Yang Zhang
Abstract: Despite being prevalent in the general field of Natural Language Processing (NLP), pre-trained language models inherently carry privacy and copyright concerns due to their nature of training on large-scale web-scraped data. In this paper, we pioneer a systematic exploration of such risks associated with pre-trained language encoders, specifically focusing on the membership leakage of pre-training data exposed through downstream models adapted from pre-trained language encoders-an aspect largely overlooked in existing literature. Our study encompasses comprehensive experiments across four types of pre-trained encoder architectures, three representative downstream tasks, and five benchmark datasets. Intriguingly, our evaluations reveal, for the first time, the existence of membership leakage even when only the black-box output of the downstream model is exposed, highlighting a privacy risk far greater than previously assumed. Alongside, we present in-depth analysis and insights toward guiding future researchers and practitioners in addressing the privacy considerations in developing pre-trained language models.
Authors: Jian Chen, Vashisth Tiwari, Ranajoy Sadhukhan, Zhuoming Chen, Jinyuan Shi, Ian En-Hsu Yen, Beidi Chen
Abstract: Large Language Models (LLMs) have become more prevalent in long-context applications such as interactive chatbots, document analysis, and agent workflows, but it is challenging to serve long-context requests with low latency and high throughput. Speculative decoding (SD) is a widely used technique to reduce latency without sacrificing performance but the conventional wisdom suggests that its efficacy is limited to small batch sizes. In MagicDec, we show that surprisingly SD can achieve speedup even for a high throughput inference regime for moderate to long sequences. More interestingly, an intelligent drafting strategy can achieve better speedup with increasing batch size based on our rigorous analysis. MagicDec first identifies the bottleneck shifts with increasing batch size and sequence length, and uses these insights to deploy speculative decoding more effectively for high throughput inference. Then, it leverages draft models with sparse KV cache to address the KV bottleneck that scales with both sequence length and batch size.
Authors: Jiajun Xu, Qun Wang, Yuhang Cao, Baitao Zeng, Sicheng Liu
Abstract: This paper investigates integrating large language models (LLMs) with advanced hardware, focusing on developing a general-purpose device designed for enhanced interaction with LLMs. Initially, we analyze the current landscape, where virtual assistants and LLMs are reshaping human-technology interactions, highlighting pivotal advancements and setting the stage for a new era of intelligent hardware. Despite substantial progress in LLM technology, a significant gap exists in hardware development, particularly concerning scalability, efficiency, affordability, and multimodal capabilities. This disparity presents both challenges and opportunities, underscoring the need for hardware that is not only powerful but also versatile and capable of managing the sophisticated demands of modern computation. Our proposed device addresses these needs by emphasizing scalability, multimodal data processing, enhanced user interaction, and privacy considerations, offering a comprehensive platform for LLM integration in various applications.
Authors: Ananya Pandey, Dinesh Kumar Vishwakarma
Abstract: Various linguistic and non-linguistic clues, such as excessive emphasis on a word, a shift in the tone of voice, or an awkward expression, frequently convey sarcasm. The computer vision problem of sarcasm recognition in conversation aims to identify hidden sarcastic, criticizing, and metaphorical information embedded in everyday dialogue. Prior, sarcasm recognition has focused mainly on text. Still, it is critical to consider all textual information, audio stream, facial expression, and body position for reliable sarcasm identification. Hence, we propose a novel approach that combines a lightweight depth attention module with a self-regulated ConvNet to concentrate on the most crucial features of visual data and an attentional tokenizer based strategy to extract the most critical context-specific information from the textual data. The following is a list of the key contributions that our experimentation has made in response to performing the task of Multi-modal Sarcasm Recognition: an attentional tokenizer branch to get beneficial features from the glossary content provided by the subtitles; a visual branch for acquiring the most prominent features from the video frames; an utterance-level feature extraction from acoustic content and a multi-headed attention based feature fusion branch to blend features obtained from multiple modalities. Extensive testing on one of the benchmark video datasets, MUSTaRD, yielded an accuracy of 79.86% for speaker dependent and 76.94% for speaker independent configuration demonstrating that our approach is superior to the existing methods. We have also conducted a cross-dataset analysis to test the adaptability of VyAnG-Net with unseen samples of another dataset MUStARD++.
Authors: Manon Revel, Matteo Cargnelutti, Tyna Eloundou, Greg Leppert
Abstract: Reinforcement Learning from Human Feedback (RLHF) aims to align language models (LMs) with human values by training reward models (RMs) on binary preferences and using these RMs to fine-tune the base LMs. Despite its importance, the internal mechanisms of RLHF remain poorly understood. This paper introduces new metrics to evaluate the effectiveness of modeling and aligning human values, namely feature imprint, alignment resistance and alignment robustness. We categorize alignment datasets into target features (desired values) and spoiler features (undesired concepts). By regressing RM scores against these features, we quantify the extent to which RMs reward them - a metric we term feature imprint. We define alignment resistance as the proportion of the preference dataset where RMs fail to match human preferences, and we assess alignment robustness by analyzing RM responses to perturbed inputs. Our experiments, utilizing open-source components like the Anthropic/hh-rlhf preference dataset and OpenAssistant RMs, reveal significant imprints of target features and a notable sensitivity to spoiler features. We observed a 26% incidence of alignment resistance in portions of the dataset where LM-labelers disagreed with human preferences. Furthermore, we find that misalignment often arises from ambiguous entries within the alignment dataset. These findings underscore the importance of scrutinizing both RMs and alignment datasets for a deeper understanding of value alignment.
Authors: Yue Zhang, Parisa Kordjamshidi
Abstract: The existing methods for Vision and Language Navigation in the Continuous Environment (VLN-CE) commonly incorporate a waypoint predictor to discretize the environment. This simplifies the navigation actions into a view selection task and improves navigation performance significantly compared to direct training using low-level actions. However, the VLN-CE agents are still far from the real robots since there are gaps between their visual perception and executed actions. First, VLN-CE agents that discretize the visual environment are primarily trained with high-level view selection, which causes them to ignore crucial spatial reasoning within the low-level action movements. Second, in these models, the existing waypoint predictors neglect object semantics and their attributes related to passibility, which can be informative in indicating the feasibility of actions. To address these two issues, we introduce a low-level action decoder jointly trained with high-level action prediction, enabling the current VLN agent to learn and ground the selected visual view to the low-level controls. Moreover, we enhance the current waypoint predictor by utilizing visual representations containing rich semantic information and explicitly masking obstacles based on humans' prior knowledge about the feasibility of actions. Empirically, our agent can improve navigation performance metrics compared to the strong baselines on both high-level and low-level actions.
Authors: Matthew Tassava, Cameron Kolodjski, Jordan Milbrath, Adorah Bishop, Nathan Flanders, Robbie Fetsch, Danielle Hanson, Jeremy Straub
Abstract: This paper presents and evaluates work on the development of an artificial intelligence (AI) anti-bullying system. The system is designed to identify coordinated bullying attacks via social media and other mechanisms, characterize them and propose remediation and response activities to them. In particular, a large language model (LLM) is used to populate an enhanced expert system-based network model of a bullying attack. This facilitates analysis and remediation activity - such as generating report messages to social media companies - determination. The system is described and the efficacy of the LLM for populating the model is analyzed herein.
Authors: Yonghui Xiao, Yuxin Ding, Changwan Ryu, Petr Zadrazil, Francoise Beaufays
Abstract: Federated learning (FL) has shown promising results on training machine learning models with privacy preservation. However, for large models with over 100 million parameters, the training resource requirement becomes an obstacle for FL because common devices do not have enough memory and computation power to finish the FL tasks. Although efficient training methods have been proposed, it is still a challenge to train the large models like Conformer based ASR. This paper presents a systematic solution to train the full-size ASR models of 130M parameters with FL. To our knowledge, this is the first real-world FL application of the Conformer model, which is also the largest model ever trained with FL so far. And this is the first paper showing FL can improve the ASR model quality with a set of proposed methods to refine the quality of data and labels of clients. We demonstrate both the training efficiency and the model quality improvement in real-world experiments.
Authors: Jinxin Liu, Zao Yang
Abstract: The responses generated by Large Language Models (LLMs) can include sensitive information from individuals and organizations, leading to potential privacy leakage. This work implements Influence Functions (IFs) to trace privacy leakage back to the training data, thereby mitigating privacy concerns of Language Models (LMs). However, we notice that current IFs struggle to accurately estimate the influence of tokens with large gradient norms, potentially overestimating their influence. When tracing the most influential samples, this leads to frequently tracing back to samples with large gradient norm tokens, overshadowing the actual most influential samples even if their influences are well estimated. To address this issue, we propose Heuristically Adjusted IF (HAIF), which reduces the weight of tokens with large gradient norms, thereby significantly improving the accuracy of tracing the most influential samples. To establish easily obtained groundtruth for tracing privacy leakage, we construct two datasets, PII-E and PII-CR, representing two distinct scenarios: one with identical text in the model outputs and pre-training data, and the other where models leverage their reasoning abilities to generate text divergent from pre-training data. HAIF significantly improves tracing accuracy, enhancing it by 20.96\% to 73.71\% on the PII-E dataset and 3.21\% to 45.93\% on the PII-CR dataset, compared to the best SOTA IFs against various GPT-2 and QWen-1.5 models. HAIF also outperforms SOTA IFs on real-world pretraining data CLUECorpus2020, demonstrating strong robustness regardless prompt and response lengths.
Authors: Xuan Xie, Jiayang Song, Yuheng Huang, Da Song, Fuyuan Zhang, Felix Juefei-Xu, Lei Ma
Abstract: Large Language Models (LLMs) are widely used in many different domains, but because of their limited interpretability, there are questions about how trustworthy they are in various perspectives, e.g., truthfulness and toxicity. Recent research has started developing testing methods for LLMs, aiming to uncover untrustworthy issues, i.e., defects, before deployment. However, systematic and formalized testing criteria are lacking, which hinders a comprehensive assessment of the extent and adequacy of testing exploration. To mitigate this threat, we propose a set of multi-level testing criteria, LeCov, for LLMs. The criteria consider three crucial LLM internal components, i.e., the attention mechanism, feed-forward neurons, and uncertainty, and contain nine types of testing criteria in total. We apply the criteria in two scenarios: test prioritization and coverage-guided testing. The experiment evaluation, on three models and four datasets, demonstrates the usefulness and effectiveness of LeCov.
Authors: Xiao Wang, Yao Rong, Fuling Wang, Jianing Li, Lin Zhu, Bo Jiang, Yaowei Wang
Abstract: Sign Language Translation (SLT) is a core task in the field of AI-assisted disability. Unlike traditional SLT based on visible light videos, which is easily affected by factors such as lighting, rapid hand movements, and privacy breaches, this paper proposes the use of high-definition Event streams for SLT, effectively mitigating the aforementioned issues. This is primarily because Event streams have a high dynamic range and dense temporal signals, which can withstand low illumination and motion blur well. Additionally, due to their sparsity in space, they effectively protect the privacy of the target person. More specifically, we propose a new high-resolution Event stream sign language dataset, termed Event-CSL, which effectively fills the data gap in this area of research. It contains 14,827 videos, 14,821 glosses, and 2,544 Chinese words in the text vocabulary. These samples are collected in a variety of indoor and outdoor scenes, encompassing multiple angles, light intensities, and camera movements. We have benchmarked existing mainstream SLT works to enable fair comparison for future efforts. Based on this dataset and several other large-scale datasets, we propose a novel baseline method that fully leverages the Mamba model's ability to integrate temporal information of CNN features, resulting in improved sign language translation outcomes. Both the benchmark dataset and source code will be released on https://github.com/Event-AHU/OpenESL
Authors: Adel Elmahdy, Sheng-Chieh Lin, Amin Ahmad
Abstract: Information retrieval across different languages is an increasingly important challenge in natural language processing. Recent approaches based on multilingual pre-trained language models have achieved remarkable success, yet they often optimize for either monolingual, cross-lingual, or multilingual retrieval performance at the expense of others. This paper proposes a novel hybrid batch training strategy to simultaneously improve zero-shot retrieval performance across monolingual, cross-lingual, and multilingual settings while mitigating language bias. The approach fine-tunes multilingual language models using a mix of monolingual and cross-lingual question-answer pair batches sampled based on dataset size. Experiments on XQuAD-R, MLQA-R, and MIRACL benchmark datasets show that the proposed method consistently achieves comparable or superior results in zero-shot retrieval across various languages and retrieval tasks compared to monolingual-only or cross-lingual-only training. Hybrid batch training also substantially reduces language bias in multilingual retrieval compared to monolingual training. These results demonstrate the effectiveness of the proposed approach for learning language-agnostic representations that enable strong zero-shot retrieval performance across diverse languages.
Authors: Yupeng Su, Ziyi Guan, Xiaoqun Liu, Tianlai Jin, Dongkuan Wu, Graziano Chesi, Ngai Wong, Hao Yu
Abstract: Large language models (LLMs) have grown significantly in scale, leading to a critical need for efficient model pruning techniques. Existing post-training pruning techniques primarily focus on measuring weight importance on converged dense models to determine salient weights to retain. However, they often overlook the changes in weight importance during the pruning process, which can lead to performance degradation in the pruned models. To address this issue, we present LLM-Barber (Block-Aware Rebuilder for Sparsity Mask in One-Shot), a novel one-shot pruning framework that rebuilds the sparsity mask of pruned models without any retraining or weight reconstruction. LLM-Barber incorporates block-aware error optimization across Self-Attention and MLP blocks, ensuring global performance optimization. Inspired by the recent discovery of prominent outliers in LLMs, LLM-Barber introduces an innovative pruning metric that identifies weight importance using weights multiplied by gradients. Our experiments show that LLM-Barber can efficiently prune models like LLaMA and OPT families with 7B to 13B parameters on a single A100 GPU in just 30 minutes, achieving state-of-the-art results in both perplexity and zero-shot performance across various language benchmarks. Code is available at https://github.com/YupengSu/LLM-Barber.
Authors: Jonathan Light, Min Cai, Weiqin Chen, Guanzhi Wang, Xiusi Chen, Wei Cheng, Yisong Yue, Ziniu Hu
Abstract: In this paper, we propose a new method Strategist that utilizes LLMs to acquire new skills for playing multi-agent games through a self-improvement process. Our method gathers quality feedback through self-play simulations with Monte Carlo tree search and LLM-based reflection, which can then be used to learn high-level strategic skills such as how to evaluate states that guide the low-level execution.We showcase how our method can be used in both action planning and dialogue generation in the context of games, achieving good performance on both tasks. Specifically, we demonstrate that our method can help train agents with better performance than both traditional reinforcement learning-based approaches and other LLM-based skill learning approaches in games including the Game of Pure Strategy (GOPS) and The Resistance: Avalon.
Authors: Shiming Xie, Hong Chen, Fred Yu, Zeye Sun, Xiuyu Wu
Abstract: Instruct LLM provide a paradigm used in large scale language model to align LLM to human preference. The paradigm contains supervised fine tuning and reinforce learning from human feedback. This paradigm is also used in downstream scenarios to adapt LLM to specific corpora and applications. Comparing to SFT, there are many efforts focused on RLHF and several algorithms being proposed, such as PPO, DPO, IPO, KTO, MinorDPO and etc. Meanwhile most efforts for SFT are focused on how to collect, filter and mix high quality data. In this article with insight from DPO and MinorDPO, we propose a training metric for SFT to measure the discrepancy between the optimized model and the original model, and a loss function MinorSFT that can increase the training effectiveness, and reduce the discrepancy between the optimized LLM and original LLM.
Authors: Yuwei Zhao, Ziyang Luo, Yuchen Tian, Hongzhan Lin, Weixiang Yan, Annan Li, Jing Ma
Abstract: Recent advancements in large language models (LLMs) have showcased impressive code generation capabilities, primarily evaluated through language-to-code benchmarks. However, these benchmarks may not fully capture a model's code understanding abilities. We introduce CodeJudge-Eval (CJ-Eval), a novel benchmark designed to assess LLMs' code understanding abilities from the perspective of code judging rather than code generation. CJ-Eval challenges models to determine the correctness of provided code solutions, encompassing various error types and compilation issues. By leveraging a diverse set of problems and a fine-grained judging system, CJ-Eval addresses the limitations of traditional benchmarks, including the potential memorization of solutions. Evaluation of 12 well-known LLMs on CJ-Eval reveals that even state-of-the-art models struggle, highlighting the benchmark's ability to probe deeper into models' code understanding abilities. Our benchmark will be available at \url{https://github.com/CodeLLM-Research/CodeJudge-Eval}.
Authors: Chenxing Wei, Yao Shu, Ying Tiffany He, Fei Richard Yu
Abstract: Large Language Models (LLMs) are driving advancements in artificial intelligence by increasing the scale of model parameters, which has significantly enhanced generalization ability and unlocked new capabilities in practice. However, their performance in specific downstream tasks is usually hindered by their knowledge boundaries on these tasks. Thus, fine-tuning techniques, especially the widely used Low-Rank Adaptation (LoRA) method, have been introduced to expand the boundaries on these tasks, whereas LoRA would underperform on certain tasks owing to its potential overfitting on these tasks. To overcome this overfitting and improve the performance of LoRA, we propose the flexible low rank adaptation (Flexora) method to automatically and flexibly select the most important layers needing to be fine-tuned to achieve the best performance on different downstream tasks. Specifically, Flexora firstly frames this layer selection problem as a well-defined hyperparameter optimization (HPO) problem, then addresses it using the unrolled differentiation (UD) method, and finally selects the most useful layers based on the optimized hyperparameters. Our extensive experiments on many pretrained models and natural language tasks show that Flexora is able to consistently improve over the existing baselines, indicating the effectiveness of our Flexora in practice. We additionally provide insightful theoretical results and many ablation studies to deliver a comprehensive understanding of our Flexora.
Authors: Yuanhao Zeng, Fei Ren, Xinpeng Zhou, Yihang Wang, Yingxia Shao
Abstract: Although instruction tuning is widely used to adjust behavior in Large Language Models (LLMs), extensive empirical evidence and research indicates that it is primarily a process where the model fits to specific task formats, rather than acquiring new knowledge or capabilities. We propose that this limitation stems from biased features learned during instruction tuning, which differ from ideal task-specfic features, leading to learn less underlying semantics in downstream tasks. However, ideal features are unknown and incalculable, constraining past work to rely on prior knowledge to assist reasoning or training, which limits LLMs' capabilities to the developers' abilities, rather than data-driven scalable learning. In our paper, through our novel data synthesis method, DELIA (Diversity-Enhanced Learning for Instruction Adaptation), we leverage the buffering effect of extensive diverse data in LLMs training to transform biased features in instruction tuning into approximations of ideal features, without explicit prior ideal features. Experiments show DELIA's better performance compared to common instruction tuning and other baselines. It outperforms common instruction tuning by 17.07%-33.41% on Icelandic-English translation bleurt score (WMT-21 dataset, gemma-7b-it) and improves accuracy by 36.1% on formatted text generation (Llama2-7b-chat). Notably, among knowledge injection methods we've known, DELIA uniquely align the internal representations of new special tokens with their prior semantics.
Authors: Yuyan Chen, Chenwei Wu, Songzhou Yan, Panjun Liu, Haoyu Zhou, Yanghua Xiao
Abstract: Teachers are important to imparting knowledge and guiding learners, and the role of large language models (LLMs) as potential educators is emerging as an important area of study. Recognizing LLMs' capability to generate educational content can lead to advances in automated and personalized learning. While LLMs have been tested for their comprehension and problem-solving skills, their capability in teaching remains largely unexplored. In teaching, questioning is a key skill that guides students to analyze, evaluate, and synthesize core concepts and principles. Therefore, our research introduces a benchmark to evaluate the questioning capability in education as a teacher of LLMs through evaluating their generated educational questions, utilizing Anderson and Krathwohl's taxonomy across general, monodisciplinary, and interdisciplinary domains. We shift the focus from LLMs as learners to LLMs as educators, assessing their teaching capability through guiding them to generate questions. We apply four metrics, including relevance, coverage, representativeness, and consistency, to evaluate the educational quality of LLMs' outputs. Our results indicate that GPT-4 demonstrates significant potential in teaching general, humanities, and science courses; Claude2 appears more apt as an interdisciplinary teacher. Furthermore, the automatic scores align with human perspectives.
Authors: Yunzhe Xu, Yiyuan Pan, Zhe Liu, Hesheng Wang
Abstract: Large Language Models (LLMs) have demonstrated potential in Vision-and-Language Navigation (VLN) tasks, yet current applications face challenges. While LLMs excel in general conversation scenarios, they struggle with specialized navigation tasks, yielding suboptimal performance compared to specialized VLN models. We introduce FLAME (FLAMingo-Architected Embodied Agent), a novel Multimodal LLM-based agent and architecture designed for urban VLN tasks that efficiently handles multiple observations. Our approach implements a three-phase tuning technique for effective adaptation to navigation tasks, including single perception tuning for street view description, multiple perception tuning for trajectory summarization, and end-to-end training on VLN datasets. The augmented datasets are synthesized automatically. Experimental results demonstrate FLAME's superiority over existing methods, surpassing state-of-the-art methods by a 7.3% increase in task completion rate on Touchdown dataset. This work showcases the potential of Multimodal LLMs (MLLMs) in complex navigation tasks, representing an advancement towards practical applications of MLLMs in embodied AI. Project page: https://flame-sjtu.github.io
Authors: Jaehun Jung, Peter West, Liwei Jiang, Faeze Brahman, Ximing Lu, Jillian Fisher, Taylor Sorensen, Yejin Choi
Abstract: We present Impossible Distillation, a novel framework for paraphrasing and sentence summarization, that distills a high-quality dataset and model from a low-quality teacher that itself cannot perform these tasks. Unlike prior works that rely on an extreme-scale teacher model (e.g., GPT3) or task-specific architecture, we hypothesize and verify the paraphrastic proximity intrinsic to pre-trained LMs (e.g., GPT2), where paraphrases occupy a proximal subspace in the LM distribution. By identifying and distilling generations from these subspaces, Impossible Distillation produces a high-quality dataset and model even from GPT2-scale LMs. We evaluate our method on multiple benchmarks spanning unconstrained / syntax-controlled paraphrase generation and sentence summarization. Our model with 770M parameters consistently outperforms strong baselines, including models distilled from ChatGPT, and sometimes, even ChatGPT itself. Also, we find that our distilled dataset from 1.5B LMs exhibits higher diversity and fidelity than up to 13 times larger datasets.
Authors: Iknoor Singh, Carolina Scarton, Xingyi Song, Kalina Bontcheva
Abstract: Finding previously debunked narratives involves identifying claims that have already undergone fact-checking. The issue intensifies when similar false claims persist in multiple languages, despite the availability of debunks for several months in another language. Hence, automatically finding debunks (or fact-checks) in multiple languages is crucial to make the best use of scarce fact-checkers' resources. Mainly due to the lack of readily available data, this is an understudied problem, particularly when considering the cross-lingual scenario, i.e. the retrieval of debunks in a language different from the language of the online post being checked. This study introduces cross-lingual debunked narrative retrieval and addresses this research gap by: (i) creating Multilingual Misinformation Tweets (MMTweets): a dataset that stands out, featuring cross-lingual pairs, images, human annotations, and fine-grained labels, making it a comprehensive resource compared to its counterparts; (ii) conducting an extensive experiment to benchmark state-of-the-art cross-lingual retrieval models and introducing multistage retrieval methods tailored for the task; and (iii) comprehensively evaluating retrieval models for their cross-lingual and cross-dataset transfer capabilities within MMTweets, and conducting a retrieval latency analysis. We find that MMTweets presents challenges for cross-lingual debunked narrative retrieval, highlighting areas for improvement in retrieval models. Nonetheless, the study provides valuable insights for creating MMTweets datasets and optimising debunked narrative retrieval models to empower fact-checking endeavours. The dataset and annotation codebook are publicly available at https://doi.org/10.5281/zenodo.10637161.
Authors: Jiaying Wu, Jiafeng Guo, Bryan Hooi
Abstract: It is commonly perceived that fake news and real news exhibit distinct writing styles, such as the use of sensationalist versus objective language. However, we emphasize that style-related features can also be exploited for style-based attacks. Notably, the advent of powerful Large Language Models (LLMs) has empowered malicious actors to mimic the style of trustworthy news sources, doing so swiftly, cost-effectively, and at scale. Our analysis reveals that LLM-camouflaged fake news content significantly undermines the effectiveness of state-of-the-art text-based detectors (up to 38% decrease in F1 Score), implying a severe vulnerability to stylistic variations. To address this, we introduce SheepDog, a style-robust fake news detector that prioritizes content over style in determining news veracity. SheepDog achieves this resilience through (1) LLM-empowered news reframings that inject style diversity into the training process by customizing articles to match different styles; (2) a style-agnostic training scheme that ensures consistent veracity predictions across style-diverse reframings; and (3) content-focused veracity attributions that distill content-centric guidelines from LLMs for debunking fake news, offering supplementary cues and potential intepretability that assist veracity prediction. Extensive experiments on three real-world benchmarks demonstrate SheepDog's style robustness and adaptability to various backbones.
Authors: Eliseo Bao, Anxo P\'erez, Javier Parapar
Abstract: Users of social platforms often perceive these sites as supportive spaces to post about their mental health issues. Those conversations contain important traces about individuals' health risks. Recently, researchers have exploited this online information to construct mental health detection models, which aim to identify users at risk on platforms like Twitter, Reddit or Facebook. Most of these models are centred on achieving good classification results, ignoring the explainability and interpretability of the decisions. Recent research has pointed out the importance of using clinical markers, such as the use of symptoms, to improve trust in the computational models by health professionals. In this paper, we propose using transformer-based architectures to detect and explain the appearance of depressive symptom markers in the users' writings. We present two approaches: i) train a model to classify, and another one to explain the classifier's decision separately and ii) unify the two tasks simultaneously using a single model. Additionally, for this latter manner, we also investigated the performance of recent conversational LLMs when using in-context learning. Our natural language explanations enable clinicians to interpret the models' decisions based on validated symptoms, enhancing trust in the automated process. We evaluate our approach using recent symptom-based datasets, employing both offline and expert-in-the-loop metrics to assess the quality of the explanations generated by our models. The experimental results show that it is possible to achieve good classification results while generating interpretable symptom-based explanations.
Authors: Yong Ma, Senlin Luo, Yu-Ming Shang, Yifei Zhang, Zhengjun Li
Abstract: Researchers have investigated the potential of leveraging pre-trained language models, such as CodeBERT, to enhance source code-related tasks. Previous methodologies have relied on CodeBERT's '[CLS]' token as the embedding representation of input sequences for task performance, necessitating additional neural network layers to enhance feature representation, which in turn increases computational expenses. These approaches have also failed to fully leverage the comprehensive knowledge inherent within the source code and its associated text, potentially limiting classification efficacy. We propose CodeClassPrompt, a text classification technique that harnesses prompt learning to extract rich knowledge associated with input sequences from pre-trained models, thereby eliminating the need for additional layers and lowering computational costs. By applying an attention mechanism, we synthesize multi-layered knowledge into task-specific features, enhancing classification accuracy. Our comprehensive experimentation across four distinct source code-related tasks reveals that CodeClassPrompt achieves competitive performance while significantly reducing computational overhead.
Authors: Zaibin Zhang, Yongting Zhang, Lijun Li, Hongzhi Gao, Lijun Wang, Huchuan Lu, Feng Zhao, Yu Qiao, Jing Shao
Abstract: Multi-agent systems, when enhanced with Large Language Models (LLMs), exhibit profound capabilities in collective intelligence. However, the potential misuse of this intelligence for malicious purposes presents significant risks. To date, comprehensive research on the safety issues associated with multi-agent systems remains limited. In this paper, we explore these concerns through the innovative lens of agent psychology, revealing that the dark psychological states of agents constitute a significant threat to safety. To tackle these concerns, we propose a comprehensive framework (PsySafe) grounded in agent psychology, focusing on three key areas: firstly, identifying how dark personality traits in agents can lead to risky behaviors; secondly, evaluating the safety of multi-agent systems from the psychological and behavioral perspectives, and thirdly, devising effective strategies to mitigate these risks. Our experiments reveal several intriguing phenomena, such as the collective dangerous behaviors among agents, agents' self-reflection when engaging in dangerous behavior, and the correlation between agents' psychological assessments and dangerous behaviors. We anticipate that our framework and observations will provide valuable insights for further research into the safety of multi-agent systems. We will make our data and code publicly accessible at https://github.com/AI4Good24/PsySafe.
Authors: Yu Sun, Keyu Chen, Shujie Wang, Peiji Li, Qipeng Guo, Hang Yan, Xipeng Qiu, Xuanjing Huang, Dahua Lin
Abstract: Large language models (LLMs) garner significant attention for their unprecedented performance, leading to an increasing number of researches evaluating LLMs. However, these evaluation benchmarks are limited to assessing the instruction-following capabilities, overlooking the fundamental abilities that emerge during the pre-training stage. Previous subjective evaluation methods mainly reply on scoring by API models. However, in the absence of references, large models have shown limited ability to discern subtle differences. To bridge the gap, we propose F-Eval, a bilingual evaluation benchmark to evaluate the fundamental abilities, including expression, commonsense and logic. The tasks in F-Eval include multi-choice objective tasks, open-ended objective tasks, reference-based subjective tasks and reference-free subjective tasks. For reference-free subjective tasks, we devise new evaluation methods, serving as alternatives to scoring by API models. We conduct evaluations on 13 advanced LLMs. Results show that our evaluation methods show higher correlation coefficients and larger distinction than other evaluators. Additionally, we discuss the influence of different model sizes, dimensions, and normalization methods. We anticipate that F-Eval will facilitate the study of LLMs' fundamental abilities.
Authors: Nicholas Asher, Swarnadeep Bhar
Abstract: Despite great performance on many tasks, language models (LMs) still struggle with reasoning, sometimes providing responses that cannot possibly be true because they stem from logical incoherence. We call such responses \textit{strong hallucinations} and prove that they follow from an LM's computation of its internal representations for logical operators and outputs from those representations. Focusing on negation, we provide a novel solution in which negation is treated not as another element of a latent representation, but as \textit{an operation over an LM's latent representations that constrains how they may evolve}. We show that our approach improves model performance in cloze prompting and natural language inference tasks with negation without requiring training on sparse negative data.
Authors: Ye Chen, Igor Couto, Wei Cai, Cong Fu, Bruno Dorneles
Abstract: We introduce SoftTiger, a clinical large language model (CLaM) designed as a foundation model for healthcare workflows. The narrative and unstructured nature of clinical notes is a major obstacle for healthcare intelligentization. We address a critical problem of structuring clinical notes into clinical data, according to international interoperability standards. We collect and annotate data for three subtasks, namely, international patient summary, clinical impression and medical encounter. We then supervised fine-tuned a state-of-the-art LLM using public and credentialed clinical data. The training is orchestrated in a way that the target model can first support basic clinical tasks such as abbreviation expansion and temporal information extraction, and then learn to perform more complex downstream clinical tasks. Moreover, we address several modeling challenges in the healthcare context, e.g., extra long context window. Our blind pairwise evaluation shows that SoftTiger outperforms other popular open-source models and GPT-3.5, comparable to Gemini-pro, with a mild gap from GPT-4. We believe that LLMs may become a step-stone towards healthcare digitalization and democratization. Therefore, we publicly release SoftTiger models at scales of 13 billion and 70 billion parameters, as well as datasets and code for our innovative scalable evaluation, hopefully, making a significant contribution to the healthcare industry.
Authors: Jaehun Jung, Ximing Lu, Liwei Jiang, Faeze Brahman, Peter West, Pang Wei Koh, Yejin Choi
Abstract: The current winning recipe for automatic summarization is using proprietary large-scale language models (LLMs) such as ChatGPT as is, or imitation learning from them as teacher models. While increasingly ubiquitous dependence on such large-scale language models is convenient, there remains an important question of whether small-scale models could have achieved competitive results, if we were to seek an alternative learning method -- that allows for a more cost-efficient, controllable, yet powerful summarizer. We present InfoSumm, a novel framework to distill a powerful summarizer based on the information-theoretic objective for summarization, without relying on either the LLM's capability or human-written references. To achieve this, we first propose a novel formulation of the desiderata of summarization (saliency, faithfulness and brevity) through the lens of mutual information between the original document and the summary. Based on this formulation, we start off from Pythia-2.8B as the teacher model, which is not yet capable of summarization, then self-train the model to optimize for the information-centric measures of ideal summaries. Distilling from the improved teacher, we arrive at a compact but powerful summarizer with only 568M parameters that performs competitively against ChatGPT, without ever relying on ChatGPT's capabilities. Extensive analysis demonstrates that our approach outperforms in-domain supervised models in human evaluation, let alone state-of-the-art unsupervised methods, and wins over ChatGPT in controllable summarization.
Authors: Amin Abolghasemi, Zhaochun Ren, Arian Askari, Mohammad Aliannejadi, Maarten de Rijke, Suzan Verberne
Abstract: An important unexplored aspect in previous work on user satisfaction estimation for Task-Oriented Dialogue (TOD) systems is their evaluation in terms of robustness for the identification of user dissatisfaction: current benchmarks for user satisfaction estimation in TOD systems are highly skewed towards dialogues for which the user is satisfied. The effect of having a more balanced set of satisfaction labels on performance is unknown. However, balancing the data with more dissatisfactory dialogue samples requires further data collection and human annotation, which is costly and time-consuming. In this work, we leverage large language models (LLMs) and unlock their ability to generate satisfaction-aware counterfactual dialogues to augment the set of original dialogues of a test collection. We gather human annotations to ensure the reliability of the generated samples. We evaluate two open-source LLMs as user satisfaction estimators on our augmented collection against state-of-the-art fine-tuned models. Our experiments show that when used as few-shot user satisfaction estimators, open-source LLMs show higher robustness to the increase in the number of dissatisfaction labels in the test collection than the fine-tuned state-of-the-art models. Our results shed light on the need for data augmentation approaches for user satisfaction estimation in TOD systems. We release our aligned counterfactual dialogues, which are curated by human annotation, to facilitate further research on this topic.
Authors: Xinpeng Wang, Chengzhi Hu, Bolei Ma, Paul R\"ottger, Barbara Plank
Abstract: Multiple choice questions (MCQs) are commonly used to evaluate the capabilities of large language models (LLMs). One common way to evaluate the model response is to rank the candidate answers based on the log probability of the first token prediction. An alternative way is to examine the text output. Prior work has shown that first token probabilities lack robustness to changes in MCQ phrasing, and that first token probabilities do not match text answers for instruction-tuned models. Therefore, in this paper, we investigate the robustness of text answers. We show that the text answers are more robust to question perturbations than the first token probabilities, when the first token answers mismatch the text answers. The difference in robustness increases as the mismatch rate becomes greater. As the mismatch reaches over 50\%, the text answer is more robust to option order changes than the debiased first token probabilities using state-of-the-art debiasing methods such as PriDe. Our findings provide further evidence for the benefits of text answer evaluation over first token probability evaluation.
Authors: Huihan Li, Liwei Jiang, Jena D. Hwang, Hyunwoo Kim, Sebastin Santy, Taylor Sorensen, Bill Yuchen Lin, Nouha Dziri, Xiang Ren, Yejin Choi
Abstract: As the utilization of large language models (LLMs) has proliferated world-wide, it is crucial for them to have adequate knowledge and fair representation for diverse global cultures. In this work, we uncover culture perceptions of three SOTA models on 110 countries and regions on 8 culture-related topics through culture-conditioned generations, and extract symbols from these generations that are associated to each culture by the LLM. We discover that culture-conditioned generation consist of linguistic "markers" that distinguish marginalized cultures apart from default cultures. We also discover that LLMs have an uneven degree of diversity in the culture symbols, and that cultures from different geographic regions have different presence in LLMs' culture-agnostic generation. Our findings promote further research in studying the knowledge and fairness of global culture perception in LLMs. Code and Data can be found here: https://github.com/huihanlhh/Culture-Gen/
Authors: Ryo Kamoi, Yusen Zhang, Nan Zhang, Jiawei Han, Rui Zhang
Abstract: Self-correction is an approach to improving responses from large language models (LLMs) by refining the responses using LLMs during inference. Prior work has proposed various self-correction frameworks using different sources of feedback, including self-evaluation and external feedback. However, there is still no consensus on the question of when LLMs can correct their own mistakes, as recent studies also report negative results. In this work, we critically survey broad papers and discuss the conditions required for successful self-correction. We first find that prior studies often do not define their research questions in detail and involve impractical frameworks or unfair evaluations that over-evaluate self-correction. To tackle these issues, we categorize research questions in self-correction research and provide a checklist for designing appropriate experiments. Our critical survey based on the newly categorized research questions shows that (1) no prior work demonstrates successful self-correction with feedback from prompted LLMs, except for studies in tasks that are exceptionally suited for self-correction, (2) self-correction works well in tasks that can use reliable external feedback, and (3) large-scale fine-tuning enables self-correction.
Authors: Hao Li, Yuping Wu, Viktor Schlegel, Riza Batista-Navarro, Tharindu Madusanka, Iqra Zahid, Jiayan Zeng, Xiaochi Wang, Xinran He, Yizhi Li, Goran Nenadic
Abstract: With the recent advances of large language models (LLMs), it is no longer infeasible to build an automated debate system that helps people to synthesise persuasive arguments. Previous work attempted this task by integrating multiple components. In our work, we introduce an argument mining dataset that captures the end-to-end process of preparing an argumentative essay for a debate, which covers the tasks of claim and evidence identification (Task 1 ED), evidence convincingness ranking (Task 2 ECR), argumentative essay summarisation and human preference ranking (Task 3 ASR) and metric learning for automated evaluation of resulting essays, based on human feedback along argument quality dimensions (Task 4 SQE). Our dataset contains 14k examples of claims that are fully annotated with the various properties supporting the aforementioned tasks. We evaluate multiple generative baselines for each of these tasks, including representative LLMs. We find, that while they show promising results on individual tasks in our benchmark, their end-to-end performance on all four tasks in succession deteriorates significantly, both in automated measures as well as in human-centred evaluation. This challenge presented by our proposed dataset motivates future research on end-to-end argument mining and summarisation. The repository of this project is available at https://github.com/HaoBytes/ArgSum-Datatset
Authors: David Wadden, Kejian Shi, Jacob Morrison, Aakanksha Naik, Shruti Singh, Nitzan Barzilay, Kyle Lo, Tom Hope, Luca Soldaini, Shannon Zejiang Shen, Doug Downey, Hannaneh Hajishirzi, Arman Cohan
Abstract: We present SciRIFF (Scientific Resource for Instruction-Following and Finetuning), a dataset of 137K instruction-following demonstrations for 54 tasks covering five essential scientific literature understanding capabilities: information extraction, summarization, question answering, claim verification, and classification. SciRIFF demonstrations are notable for their long input contexts, detailed task specifications, and complex structured outputs. While instruction-following resources are available in specific domains such as clinical medicine and chemistry, SciRIFF is the first dataset focused on extracting and synthesizing information from research literature across a wide range of scientific fields. To demonstrate the utility of SciRIFF, we develop a sample-efficient strategy to adapt a general instruction-following model for science by performing additional finetuning on a mix of general-domain and SciRIFF demonstrations. In evaluations on nine held-out scientific tasks, our model -- called SciTulu -- improves over a strong LLM baseline by 28.1% and 6.5% at the 7B and 70B scales respectively, while maintaining general instruction-following performance within 2% of the baseline. We are optimistic that SciRIFF will facilitate the development and evaluation of LLMs to help researchers navigate the ever-growing body of scientific literature. We release our dataset, model checkpoints, and data processing and evaluation code to enable further research.
Authors: Lihu Chen, Adam Dejl, Francesca Toni
Abstract: Large Language Models (LLMs) possess vast amounts of knowledge within their parameters, prompting research into methods for locating and editing this knowledge. Previous work has largely focused on locating entity-related (often single-token) facts in smaller models. However, several key questions remain unanswered: (1) How can we effectively locate query-relevant neurons in contemporary autoregressive LLMs, such as Llama and Mistral? (2) How can we address the challenge of long-form text generation? (3) Are there localized knowledge regions in LLMs? In this study, we introduce Query-Relevant Neuron Cluster Attribution (QRNCA), a novel architecture-agnostic framework capable of identifying query-relevant neurons in LLMs. QRNCA allows for the examination of long-form answers beyond triplet facts by employing the proxy task of multi-choice question answering. To evaluate the effectiveness of our detected neurons, we build two multi-choice QA datasets spanning diverse domains and languages. Empirical evaluations demonstrate that our method outperforms baseline methods significantly. Further, analysis of neuron distributions reveals the presence of visible localized regions, particularly within different domains. Finally, we show potential applications of our detected neurons in knowledge editing and neuron-based prediction.
Authors: Priyanka Kargupta, Ishika Agarwal, Dilek Hakkani-Tur, Jiawei Han
Abstract: Socratic questioning is an effective teaching strategy, encouraging critical thinking and problem-solving. The conversational capabilities of large language models (LLMs) show great potential for providing scalable, real-time student guidance. However, current LLMs often give away solutions directly, making them ineffective instructors. We tackle this issue in the code debugging domain with TreeInstruct, an Instructor agent guided by a novel state space-based planning algorithm. TreeInstruct asks probing questions to help students independently identify and resolve errors. It estimates a student's conceptual and syntactical knowledge to dynamically construct a question tree based on their responses and current knowledge state, effectively addressing both independent and dependent mistakes concurrently in a multi-turn interaction setting. In addition to using an existing single-bug debugging benchmark, we construct a more challenging multi-bug dataset of 150 coding problems, incorrect solutions, and bug fixes -- all carefully constructed and annotated by experts. Extensive evaluation shows TreeInstruct's state-of-the-art performance on both datasets, proving it to be a more effective instructor than baselines. Furthermore, a real-world case study with five students of varying skill levels further demonstrates TreeInstruct's ability to guide students to debug their code efficiently with minimal turns and highly Socratic questioning. We provide our code and datasets at http://github.com/agarwalishika/TreeInstruct .
Authors: Zhepei Wei, Wei-Lin Chen, Yu Meng
Abstract: Retrieval-augmented generation (RAG) has shown promising potential to enhance the accuracy and factuality of language models (LMs). However, imperfect retrievers or noisy corpora can introduce misleading or even erroneous information to the retrieved contents, posing a significant challenge to the generation quality. Existing RAG methods typically address this challenge by directly predicting final answers despite potentially noisy inputs, resulting in an implicit denoising process that is difficult to interpret and verify. On the other hand, the acquisition of explicit denoising supervision is often costly, involving significant human efforts. In this work, we propose InstructRAG, where LMs explicitly learn the denoising process through self-synthesized rationales -- First, we instruct the LM to explain how the ground-truth answer is derived from retrieved documents. Then, these rationales can be used either as demonstrations for in-context learning of explicit denoising or as supervised fine-tuning data to train the model. Compared to standard RAG approaches, InstructRAG requires no additional supervision, allows for easier verification of the predicted answers, and effectively improves generation accuracy. Experiments show InstructRAG consistently outperforms existing RAG methods in both training-free and trainable scenarios, achieving a relative improvement of 8.3% over the best baseline method on average across five knowledge-intensive benchmarks. Extensive analysis indicates that InstructRAG scales well with increased numbers of retrieved documents and consistently exhibits robust denoising ability even in out-of-domain datasets, demonstrating strong generalizability.
Authors: Jafar Isbarov, Kavsar Huseynova, Elvin Mammadov, Mammad Hajili, Duygu Ataman
Abstract: The emergence of multilingual large language models has enabled the development of language understanding and generation systems in Azerbaijani. However, most of the production-grade systems rely on cloud solutions, such as GPT-4. While there have been several attempts to develop open foundation models for Azerbaijani, these works have not found their way into common use due to a lack of systemic benchmarking. This paper encompasses several lines of work that promote open-source foundation models for Azerbaijani. We introduce (1) a large text corpus for Azerbaijani, (2) a family of encoder-only language models trained on this dataset, (3) labeled datasets for evaluating these models, and (4) extensive evaluation that covers all major open-source models with Azerbaijani support.
Authors: Taeho Hwang, Soyeong Jeong, Sukmin Cho, SeungYoon Han, Jong C. Park
Abstract: Recent advancements in Large Language Models (LLMs) have significantly improved their performance across various Natural Language Processing (NLP) tasks. However, LLMs still struggle with generating non-factual responses due to limitations in their parametric memory. Retrieval-Augmented Generation (RAG) systems address this issue by incorporating external knowledge with a retrieval module. Despite their successes, however, current RAG systems face challenges with retrieval failures and the limited ability of LLMs to filter out irrelevant information. Therefore, in this work, we propose DSLR (Document Refinement with Sentence-Level Re-ranking and Reconstruction), an unsupervised framework that decomposes retrieved documents into sentences, filters out irrelevant sentences, and reconstructs them again into coherent passages. We experimentally validate DSLR on multiple open-domain QA datasets and the results demonstrate that DSLR significantly enhances the RAG performance over conventional fixed-size passage. Furthermore, our DSLR enhances performance in specific, yet realistic scenarios without the need for additional training, providing an effective and efficient solution for refining retrieved documents in RAG systems.
Authors: Ye Jiang, Yimin Wang
Abstract: Large visual-language models (LVLMs) exhibit exceptional performance in visual-language reasoning across diverse cross-modal benchmarks. Despite these advances, recent research indicates that Large Language Models (LLMs), like GPT-3.5-turbo, underachieve compared to well-trained smaller models, such as BERT, in Fake News Detection (FND), prompting inquiries into LVLMs' efficacy in FND tasks. Although performance could improve through fine-tuning LVLMs, the substantial parameters and requisite pre-trained weights render it a resource-heavy endeavor for FND applications. This paper initially assesses the FND capabilities of two notable LVLMs, CogVLM and GPT4V, in comparison to a smaller yet adeptly trained CLIP model in a zero-shot context. The findings demonstrate that LVLMs can attain performance competitive with that of the smaller model. Next, we integrate standard in-context learning (ICL) with LVLMs, noting improvements in FND performance, though limited in scope and consistency. To address this, we introduce the \textbf{I}n-context \textbf{M}ultimodal \textbf{F}ake \textbf{N}ews \textbf{D}etection (IMFND) framework, enriching in-context examples and test inputs with predictions and corresponding probabilities from a well-trained smaller model. This strategic integration directs the LVLMs' focus towards news segments associated with higher probabilities, thereby improving their analytical accuracy. The experimental results suggest that the IMFND framework significantly boosts the FND efficiency of LVLMs, achieving enhanced accuracy over the standard ICL approach across three publicly available FND datasets.
Authors: Prannaya Gupta, Le Qi Yau, Hao Han Low, I-Shiang Lee, Hugo Maximus Lim, Yu Xin Teoh, Jia Hng Koh, Dar Win Liew, Rishabh Bhardwaj, Rajat Bhardwaj, Soujanya Poria
Abstract: WalledEval is a comprehensive AI safety testing toolkit designed to evaluate large language models (LLMs). It accommodates a diverse range of models, including both open-weight and API-based ones, and features over 35 safety benchmarks covering areas such as multilingual safety, exaggerated safety, and prompt injections. The framework supports both LLM and judge benchmarking and incorporates custom mutators to test safety against various text-style mutations, such as future tense and paraphrasing. Additionally, WalledEval introduces WalledGuard, a new, small, and performant content moderation tool, and two datasets: SGXSTest and HIXSTest, which serve as benchmarks for assessing the exaggerated safety of LLMs and judges in cultural contexts. We make WalledEval publicly available at https://github.com/walledai/walledeval.
Authors: Yiqun Zhang, Xiaocui Yang, Shi Feng, Daling Wang, Yifei Zhang, Kaisong Song
Abstract: Competitive debate is a complex task of computational argumentation. Large Language Models (LLMs) suffer from hallucinations and lack competitiveness in this field. To address these challenges, we introduce Agent for Debate (Agent4Debate), a dynamic multi-agent framework based on LLMs designed to enhance their capabilities in competitive debate. Drawing inspiration from human behavior in debate preparation and execution, Agent4Debate employs a collaborative architecture where four specialized agents, involving Searcher, Analyzer, Writer, and Reviewer, dynamically interact and cooperate. These agents work throughout the debate process, covering multiple stages from initial research and argument formulation to rebuttal and summary. To comprehensively evaluate framework performance, we construct the Competitive Debate Arena, comprising 66 carefully selected Chinese debate motions. We recruit ten experienced human debaters and collect records of 200 debates involving Agent4Debate, baseline models, and humans. The evaluation employs the Debatrix automatic scoring system and professional human reviewers based on the established Debatrix-Elo and Human-Elo ranking. Experimental results indicate that the state-of-the-art Agent4Debate exhibits capabilities comparable to those of humans. Furthermore, ablation studies demonstrate the effectiveness of each component in the agent structure.
Authors: Haohan Yuan, Siu Cheung Hui, Haopeng Zhang
Abstract: Biomedical Event Extraction (BEE) is a challenging task that involves modeling complex relationships between fine-grained entities in biomedical text. BEE has traditionally been formulated as a classification problem. With the recent technological advancements in large language models (LLMs), generation-based models that cast event extraction as a sequence generation problem have attracted much attention from the NLP research communities. However, current generative models often overlook the importance of cross-instance information from complex event structures such as nested events and overlapping events, which contribute to over 20% of the events in the benchmark datasets. In this paper, we propose an event structure-aware generative model named GenBEE, which can capture complex event structures in biomedical text for biomedical event extraction. In particular, GenBEE constructs event prompts that distill knowledge from LLMs for incorporating both label semantics and argument dependency relationships into the proposed model. In addition, GenBEE also generates prefixes with event structural prompts to incorporate structural features for improving the model's overall performance. We have evaluated the proposed GenBEE model on three widely used biomedical event extraction benchmark datasets, namely MLEE, GE11, and PHEE. Experimental results show that GenBEE has achieved state-of-the-art performance on the MLEE and GE11 datasets, and achieved competitive results when compared to the state-of-the-art classification-based models on the PHEE dataset.
Authors: Quanyu Long, Jianda Chen, Wenya Wang, Sinno Jialin Pan
Abstract: In-context learning (ICL) has proven to be a significant capability with the advancement of Large Language models (LLMs). By instructing LLMs using few-shot demonstrative examples, ICL enables them to perform a wide range of tasks without needing to update millions of parameters. This paper presents a unified framework for LLMs that allows them to self-select influential in-context examples to compose their contexts; self-rank candidates with different demonstration compositions; self-optimize the demonstration selection and ordering through reinforcement learning. Specifically, our method designs a parameter-efficient retrieval head that generates the optimized demonstration after training with rewards from LLM's own preference. Experimental results validate the proposed method's effectiveness in enhancing ICL performance. Additionally, our approach effectively identifies and selects the most representative examples for the current task, and includes more diversity in retrieval.
Authors: Chenming Tang, Zhixiang Wang, Yunfang Wu
Abstract: With the help of in-context learning (ICL), large language models (LLMs) have achieved impressive performance across various tasks. However, the function of descriptive instructions during ICL remains under-explored. In this work, we propose an ensemble prompt framework to describe the selection criteria of multiple in-context examples, and preliminary experiments on machine translation (MT) across six translation directions confirm that this framework boosts ICL perfromance. But to our surprise, LLMs might not necessarily care what the descriptions actually say, and the performance gain is primarily caused by the ensemble format, since the framework could lead to improvement even with random descriptive nouns. We further apply this new ensemble prompt on a range of commonsense, math, logical reasoning and hallucination tasks with three LLMs and achieve promising results, suggesting again that designing a proper prompt format would be much more effective and efficient than paying effort into specific descriptions. Our code will be publicly available once this paper is published.
Authors: Sher Badshah, Hassan Sajjad
Abstract: The emergence of Large Language Models (LLMs) as chat assistants capable of generating human-like conversations has amplified the need for robust evaluation methods, particularly for open-ended tasks. Conventional metrics like BLEU and ROUGE, while useful, are increasingly inadequate for capturing the subtle semantics and contextual richness of such generative outputs. We propose a reference-guided verdict method that automates the evaluation process by leveraging multiple LLMs-as-judges. Through experiments on three open-ended question-answering tasks, we demonstrate that combining multiple LLMs-as-judges significantly improves the reliability and accuracy of evaluations, particularly in complex tasks where a single model might struggle. Our findings reveal a strong correlation with human evaluations, establishing our method as a viable and effective alternative to traditional metrics and human judgments, particularly in the context of LLM-based chat assistants where the complexity and diversity of responses challenge existing benchmarks.
Authors: Jonathan Tonglet, Marie-Francine Moens, Iryna Gurevych
Abstract: To assist human fact-checkers, researchers have developed automated approaches for visual misinformation detection. These methods assign veracity scores by identifying inconsistencies between the image and its caption, or by detecting forgeries in the image. However, they neglect a crucial point of the human fact-checking process: identifying the original meta-context of the image. By explaining what is actually true about the image, fact-checkers can better detect misinformation, focus their efforts on check-worthy visual content, engage in counter-messaging before misinformation spreads widely, and make their explanation more convincing. Here, we fill this gap by introducing the task of automated image contextualization. We create 5Pils, a dataset of 1,676 fact-checked images with question-answer pairs about their original meta-context. Annotations are based on the 5 Pillars fact-checking framework. We implement a first baseline that grounds the image in its original meta-context using the content of the image and textual evidence retrieved from the open web. Our experiments show promising results while highlighting several open challenges in retrieval and reasoning. We make our code and data publicly available.
Authors: Emre K{\i}c{\i}man, Robert Ness, Amit Sharma, Chenhao Tan
Abstract: The causal capabilities of large language models (LLMs) are a matter of significant debate, with critical implications for the use of LLMs in societally impactful domains such as medicine, science, law, and policy. We conduct a "behavorial" study of LLMs to benchmark their capability in generating causal arguments. Across a wide range of tasks, we find that LLMs can generate text corresponding to correct causal arguments with high probability, surpassing the best-performing existing methods. Algorithms based on GPT-3.5 and 4 outperform existing algorithms on a pairwise causal discovery task (97%, 13 points gain), counterfactual reasoning task (92%, 20 points gain) and event causality (86% accuracy in determining necessary and sufficient causes in vignettes). We perform robustness checks across tasks and show that the capabilities cannot be explained by dataset memorization alone, especially since LLMs generalize to novel datasets that were created after the training cutoff date. That said, LLMs exhibit unpredictable failure modes, and we discuss the kinds of errors that may be improved and what are the fundamental limits of LLM-based answers. Overall, by operating on the text metadata, LLMs bring capabilities so far understood to be restricted to humans, such as using collected knowledge to generate causal graphs or identifying background causal context from natural language. As a result, LLMs may be used by human domain experts to save effort in setting up a causal analysis, one of the biggest impediments to the widespread adoption of causal methods. Given that LLMs ignore the actual data, our results also point to a fruitful research direction of developing algorithms that combine LLMs with existing causal techniques. Code and datasets are available at https://github.com/py-why/pywhy-llm.
Authors: Yuan Liu, Haodong Duan, Yuanhan Zhang, Bo Li, Songyang Zhang, Wangbo Zhao, Yike Yuan, Jiaqi Wang, Conghui He, Ziwei Liu, Kai Chen, Dahua Lin
Abstract: Large vision-language models (VLMs) have recently achieved remarkable progress, exhibiting impressive multimodal perception and reasoning abilities. However, effectively evaluating these large VLMs remains a major challenge, hindering future development in this domain. Traditional benchmarks like VQAv2 or COCO Caption provide quantitative performance measurements but lack fine-grained ability assessment and robust evaluation metrics. Meanwhile, subjective benchmarks, such as OwlEval, offer comprehensive evaluations of a model's abilities by incorporating human labor, which is not scalable and may display significant bias. In response to these challenges, we propose MMBench, a bilingual benchmark for assessing the multi-modal capabilities of VLMs. MMBench methodically develops a comprehensive evaluation pipeline, primarily comprised of the following key features: 1. MMBench is meticulously curated with well-designed quality control schemes, surpassing existing similar benchmarks in terms of the number and variety of evaluation questions and abilities; 2. MMBench introduces a rigorous CircularEval strategy and incorporates large language models to convert free-form predictions into pre-defined choices, which helps to yield accurate evaluation results for models with limited instruction-following capabilities. 3. MMBench incorporates multiple-choice questions in both English and Chinese versions, enabling an apples-to-apples comparison of VLMs' performance under a bilingual context. To summarize, MMBench is a systematically designed objective benchmark for a robust and holistic evaluation of vision-language models. We hope MMBench will assist the research community in better evaluating their models and facilitate future progress in this area. The evalutation code of MMBench has been integrated into VLMEvalKit: https://github.com/open-compass/VLMEvalKit.
Authors: Xin Qi, Xiaopeng Wang, Zhiyong Wang, Wang Liu, Mingming Ding, Shuchen Shi
Abstract: This paper presents a French text-to-speech synthesis system for the Blizzard Challenge 2023. The challenge consists of two tasks: generating high-quality speech from female speakers and generating speech that closely resembles specific individuals. Regarding the competition data, we conducted a screening process to remove missing or erroneous text data. We organized all symbols except for phonemes and eliminated symbols that had no pronunciation or zero duration. Additionally, we added word boundary and start/end symbols to the text, which we have found to improve speech quality based on our previous experience. For the Spoke task, we performed data augmentation according to the competition rules. We used an open-source G2P model to transcribe the French texts into phonemes. As the G2P model uses the International Phonetic Alphabet (IPA), we applied the same transcription process to the provided competition data for standardization. However, due to compiler limitations in recognizing special symbols from the IPA chart, we followed the rules to convert all phonemes into the phonetic scheme used in the competition data. Finally, we resampled all competition audio to a uniform sampling rate of 16 kHz. We employed a VITS-based acoustic model with the hifigan vocoder. For the Spoke task, we trained a multi-speaker model and incorporated speaker information into the duration predictor, vocoder, and flow layers of the model. The evaluation results of our system showed a quality MOS score of 3.6 for the Hub task and 3.4 for the Spoke task, placing our system at an average level among all participating teams.
Authors: Jinghan Yang, Shuming Ma, Furu Wei
Abstract: With in-context learning ability, the performance of large language models can be significantly boosted when provided with appropriate context. However, existing in-context learning methods mainly rely on human-provided contexts, such as labeled examples and explicit instructions. Writing context by humans is labor-intensive on various tasks and limits the model to tasks manageable by humans. To overcome these limitations, we propose Automatic In-Context Learning framework that enables the model to autonomously generate examples and instructions for problem-solving. With experiments across various models and datasets, results show that model-generated contexts outperform human-annotated contexts, including Few-Shot and Few-Shot-CoT methods, and surpass existing self-generated context methods like Zero-CoT and Auto-CoT.
Authors: Kaijie Zhu, Qinlin Zhao, Hao Chen, Jindong Wang, Xing Xie
Abstract: The evaluation of large language models (LLMs) is crucial to assess their performance and mitigate potential security risks. In this paper, we introduce PromptBench, a unified library to evaluate LLMs. It consists of several key components that are easily used and extended by researchers: prompt construction, prompt engineering, dataset and model loading, adversarial prompt attack, dynamic evaluation protocols, and analysis tools. PromptBench is designed to be an open, general, and flexible codebase for research purposes that can facilitate original study in creating new benchmarks, deploying downstream applications, and designing new evaluation protocols. The code is available at: https://github.com/microsoft/promptbench and will be continuously supported.
Authors: Wenxuan Wang, Haonan Bai, Jen-tse Huang, Yuxuan Wan, Youliang Yuan, Haoyi Qiu, Nanyun Peng, Michael R. Lyu
Abstract: Image generation models can generate or edit images from a given text. Recent advancements in image generation technology, exemplified by DALL-E and Midjourney, have been groundbreaking. These advanced models, despite their impressive capabilities, are often trained on massive Internet datasets, making them susceptible to generating content that perpetuates social stereotypes and biases, which can lead to severe consequences. Prior research on assessing bias within image generation models suffers from several shortcomings, including limited accuracy, reliance on extensive human labor, and lack of comprehensive analysis. In this paper, we propose BiasPainter, a novel evaluation framework that can accurately, automatically and comprehensively trigger social bias in image generation models. BiasPainter uses a diverse range of seed images of individuals and prompts the image generation models to edit these images using gender, race, and age-neutral queries. These queries span 62 professions, 39 activities, 57 types of objects, and 70 personality traits. The framework then compares the edited images to the original seed images, focusing on the significant changes related to gender, race, and age. BiasPainter adopts a key insight that these characteristics should not be modified when subjected to neutral prompts. Built upon this design, BiasPainter can trigger the social bias and evaluate the fairness of image generation models. We use BiasPainter to evaluate six widely-used image generation models, such as stable diffusion and Midjourney. Experimental results show that BiasPainter can successfully trigger social bias in image generation models. According to our human evaluation, BiasPainter can achieve 90.8% accuracy on automatic bias detection, which is significantly higher than the results reported in previous work.
Authors: Luxi He, Mengzhou Xia, Peter Henderson
Abstract: Current Large Language Models (LLMs), even those tuned for safety and alignment, are susceptible to jailbreaking. Some have found that just further fine-tuning an aligned model with benign data (i.e., data without harmful content) surprisingly leads to substantial degradation in safety. We delve into the data-centric aspects of why benign fine-tuning inadvertently contributes to jailbreaking. First, we represent fine-tuning data through two lenses: representation and gradient spaces. Additionally, we propose a bi-directional anchoring method that, during the selection process, prioritizes data points that are close to harmful examples and far from benign ones. Our approach effectively identifies subsets of benign data that are more likely to degrade the model's safety after fine-tuning. Training on just 100 of these seemingly benign datapoints surprisingly leads to the fine-tuned model affirmatively responding to >70% of tested harmful requests, compared to <20% after fine-tuning on randomly selected data. We also observe that the selected data frequently appear as lists, bullet points, or math questions, indicating a systematic pattern in fine-tuning data that contributes to jailbreaking.
Authors: Sebastian Bordt, Harsha Nori, Vanessa Rodrigues, Besmira Nushi, Rich Caruana
Abstract: While many have shown how Large Language Models (LLMs) can be applied to a diverse set of tasks, the critical issues of data contamination and memorization are often glossed over. In this work, we address this concern for tabular data. Specifically, we introduce a variety of different techniques to assess whether a language model has seen a tabular dataset during training. This investigation reveals that LLMs have memorized many popular tabular datasets verbatim. We then compare the few-shot learning performance of LLMs on datasets that were seen during training to the performance on datasets released after training. We find that LLMs perform better on datasets seen during training, indicating that memorization leads to overfitting. At the same time, LLMs show non-trivial performance on novel datasets and are surprisingly robust to data transformations. We then investigate the in-context statistical learning abilities of LLMs. While LLMs are significantly better than random at solving statistical classification problems, the sample efficiency of few-shot learning lags behind traditional statistical learning algorithms, especially as the dimension of the problem increases. This suggests that much of the observed few-shot performance on novel real-world datasets is due to the LLM's world knowledge. Overall, our results highlight the importance of testing whether an LLM has seen an evaluation dataset during pre-training. We release the https://github.com/interpretml/LLM-Tabular-Memorization-Checker Python package to test LLMs for memorization of tabular datasets.
URLs: https://github.com/interpretml/LLM-Tabular-Memorization-Checker
Authors: David Eric Austin, Anton Korikov, Armin Toroghi, Scott Sanner
Abstract: Designing preference elicitation (PE) methodologies that can quickly ascertain a user's top item preferences in a cold-start setting is a key challenge for building effective and personalized conversational recommendation (ConvRec) systems. While large language models (LLMs) enable fully natural language (NL) PE dialogues, we hypothesize that monolithic LLM NL-PE approaches lack the multi-turn, decision-theoretic reasoning required to effectively balance the exploration and exploitation of user preferences towards an arbitrary item set. In contrast, traditional Bayesian optimization PE methods define theoretically optimal PE strategies, but cannot generate arbitrary NL queries or reason over content in NL item descriptions -- requiring users to express preferences via ratings or comparisons of unfamiliar items. To overcome the limitations of both approaches, we formulate NL-PE in a Bayesian Optimization (BO) framework that seeks to actively elicit NL feedback to identify the best recommendation. Key challenges in generalizing BO to deal with natural language feedback include determining: (a) how to leverage LLMs to model the likelihood of NL preference feedback as a function of item utilities, and (b) how to design an acquisition function for NL BO that can elicit preferences in the infinite space of language. We demonstrate our framework in a novel NL-PE algorithm, PEBOL, which uses: 1) Natural Language Inference (NLI) between user preference utterances and NL item descriptions to maintain Bayesian preference beliefs, and 2) BO strategies such as Thompson Sampling (TS) and Upper Confidence Bound (UCB) to steer LLM query generation. We numerically evaluate our methods in controlled simulations, finding that after 10 turns of dialogue, PEBOL can achieve an MRR@10 of up to 0.27 compared to the best monolithic LLM baseline's MRR@10 of 0.17, despite relying on earlier and smaller LLMs.
Authors: Xiaoqian Liu, Guoqiang Hu, Yangfan Du, Erfeng He, Yingfeng Luo, Chen Xu, Tong Xiao, Jingbo Zhu
Abstract: Simultaneous speech translation (SimulST) is a demanding task that involves generating translations in real-time while continuously processing speech input. This paper offers a comprehensive overview of the recent developments in SimulST research, focusing on four major challenges. Firstly, the complexities associated with processing lengthy and continuous speech streams pose significant hurdles. Secondly, satisfying real-time requirements presents inherent difficulties due to the need for immediate translation output. Thirdly, striking a balance between translation quality and latency constraints remains a critical challenge. Finally, the scarcity of annotated data adds another layer of complexity to the task. Through our exploration of these challenges and the proposed solutions, we aim to provide valuable insights into the current landscape of SimulST research and suggest promising directions for future exploration.
Authors: Jiaqi Chen, Bingqian Lin, Xinmin Liu, Lin Ma, Xiaodan Liang, Kwan-Yee K. Wong
Abstract: LLM-based agents have demonstrated impressive zero-shot performance in vision-language navigation (VLN) task. However, existing LLM-based methods often focus only on solving high-level task planning by selecting nodes in predefined navigation graphs for movements, overlooking low-level control in navigation scenarios. To bridge this gap, we propose AO-Planner, a novel Affordances-Oriented Planner for continuous VLN task. Our AO-Planner integrates various foundation models to achieve affordances-oriented low-level motion planning and high-level decision-making, both performed in a zero-shot setting. Specifically, we employ a Visual Affordances Prompting (VAP) approach, where the visible ground is segmented by SAM to provide navigational affordances, based on which the LLM selects potential candidate waypoints and plans low-level paths towards selected waypoints. We further propose a high-level PathAgent which marks planned paths into the image input and reasons the most probable path by comprehending all environmental information. Finally, we convert the selected path into 3D coordinates using camera intrinsic parameters and depth information, avoiding challenging 3D predictions for LLMs. Experiments on the challenging R2R-CE and RxR-CE datasets show that AO-Planner achieves state-of-the-art zero-shot performance (8.8% improvement on SPL). Our method can also serve as a data annotator to obtain pseudo-labels, distilling its waypoint prediction ability into a learning-based predictor. This new predictor does not require any waypoint data from the simulator and achieves 47% SR competing with supervised methods. We establish an effective connection between LLM and 3D world, presenting novel prospects for employing foundation models in low-level motion control.
Authors: Hsiu-Yuan Huang, Zichen Wu, Yutong Yang, Junzhao Zhang, Yunfang Wu
Abstract: Nowadays, Large Language Models (LLMs) have demonstrated exceptional performance across various downstream tasks. However, it is challenging for users to discern whether the responses are generated with certainty or are fabricated to meet user expectations. Estimating the uncertainty of LLMs is particularly challenging due to their vast scale and the lack of white-box access. In this work, we propose a novel Uncertainty Tripartite Testing Paradigm (Unc-TTP) to classify LLM uncertainty, via evaluating the consistency of LLM outputs when incorporating label interference into the sampling-based approach. Based on Unc-TTP outputs, we aggregate instances into certain and uncertain categories. Further, we conduct a detailed analysis of the uncertainty properties of LLMs and show Unc-TTP's superiority over the existing sampling-based methods. In addition, we leverage the obtained uncertainty information to guide in-context example selection, demonstrating that Unc-TTP obviously outperforms retrieval-based and sampling-based approaches in selecting more informative examples. Our work paves a new way to classify the uncertainty of both open- and closed-source LLMs, and introduces a practical approach to exploit this uncertainty to improve LLMs performance.
Authors: Zitian Gao, Yihao Xiao
Abstract: In the Venture Capital(VC) industry, predicting the success of startups is challenging due to limited financial data and the need for subjective revenue forecasts. Previous methods based on time series analysis or deep learning often fall short as they fail to incorporate crucial inter-company relationships such as competition and collaboration. Regarding the issues, we propose a novel approach using GrahphRAG augmented time series model. With GraphRAG, time series predictive methods are enhanced by integrating these vital relationships into the analysis framework, allowing for a more dynamic understanding of the startup ecosystem in venture capital. Our experimental results demonstrate that our model significantly outperforms previous models in startup success predictions. To the best of our knowledge, our work is the first application work of GraphRAG.
Authors: Chi-Heng Lin, Shangqian Gao, James Seale Smith, Abhishek Patel, Shikhar Tuli, Yilin Shen, Hongxia Jin, Yen-Chang Hsu
Abstract: Large Language Models (LLMs) have reshaped the landscape of artificial intelligence by demonstrating exceptional performance across various tasks. However, substantial computational requirements make their deployment challenging on devices with limited resources. Recently, compression methods using low-rank matrix techniques have shown promise, yet these often lead to degraded accuracy or introduce significant overhead in parameters and inference latency. This paper introduces \textbf{Mo}dular \textbf{De}composition (MoDeGPT), a novel structured compression framework that does not need recovery fine-tuning while resolving the above drawbacks. MoDeGPT partitions the Transformer block into modules comprised of matrix pairs and reduces the hidden dimensions via reconstructing the module-level outputs. MoDeGPT is developed based on a theoretical framework that utilizes three well-established matrix decomposition algorithms -- Nystr\"om approximation, CR decomposition, and SVD -- and applies them to our redefined transformer modules. Our comprehensive experiments show MoDeGPT, without backward propagation, matches or surpasses previous structured compression methods that rely on gradient information, and saves 98% of compute costs on compressing a 13B model. On \textsc{Llama}-2/3 and OPT models, MoDeGPT maintains 90-95% zero-shot performance with 25-30% compression rates. Moreover, the compression can be done on a single GPU within a few hours and increases the inference throughput by up to 46%.
Authors: Rachel M. Harrison
Abstract: Random Number Generation Tasks (RNGTs) are used in psychology for examining how humans generate sequences devoid of predictable patterns. By adapting an existing human RNGT for an LLM-compatible environment, this preliminary study tests whether ChatGPT-3.5, a large language model (LLM) trained on human-generated text, exhibits human-like cognitive biases when generating random number sequences. Initial findings indicate that ChatGPT-3.5 more effectively avoids repetitive and sequential patterns compared to humans, with notably lower repeat frequencies and adjacent number frequencies. Continued research into different models, parameters, and prompting methodologies will deepen our understanding of how LLMs can more closely mimic human random generation behaviors, while also broadening their applications in cognitive and behavioral science research.
Authors: Fuzhao Xue, Yukang Chen, Dacheng Li, Qinghao Hu, Ligeng Zhu, Xiuyu Li, Yunhao Fang, Haotian Tang, Shang Yang, Zhijian Liu, Ethan He, Hongxu Yin, Pavlo Molchanov, Jan Kautz, Linxi Fan, Yuke Zhu, Yao Lu, Song Han
Abstract: Long-context capability is critical for multi-modal foundation models. We introduce LongVILA, a full-stack solution for long-context vision-language models, including system, model training, and dataset development. On the system side, we introduce the first long-context Multi-Modal Sequence Parallelism (MM-SP) system that enables long training and inference, enabling 2M context length training on 256 GPUs without any gradient checkpointing. MM-SP is 2.1x - 5.7x faster than ring sequence parallelism and 1.1x - 1.4x faster than Megatron context parallelism + tensor parallelism in text-only settings. Moreover, it seamlessly integrates with Hugging Face Transformers. For model training, we propose a five-stage pipeline comprising alignment, pre-training, short supervised fine-tuning, context extension, and long supervised fine-tuning. On datasets, we construct large-scale visual language pre-training datasets and long video instruction-following datasets to support our multi-stage training process. LongVILA extends the number of frames of VILA from 8 to 1024, and improves the long video captioning score from 2.00 to 3.26 (1.6x), achieving 99.5% accuracy in 1400-frames video (274k context length) needle-in-a-haystack. LongVILA-8B demonstrates consistent accuracy improvements on long videos in the VideoMME benchmark as the number of frames increases.