CS researchers presented their work at the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP 2024), showcasing research across natural language processing. Held from November 12-16 in Miami, this event includes diverse sessions covering topics from core NLP advancements to innovative applications. The accepted papers promise insights into cutting-edge techniques, inviting academics, practitioners, and enthusiasts to explore the latest in language processing research.
STORYSUMM: Evaluating Faithfulness in Story Summarization
Melanie Subbiah Columbia University, Faisal Ladhak Answer.AI, Akankshya Mishra Columbia University, Griffin Thomas Adams Answer.AI, Lydia Chilton Columbia University, Kathleen McKeown Columbia University
Abstract:
Human evaluation has been the gold standard for checking faithfulness in abstractive summarization. However, with a challenging source domain like narrative, multiple annotators can agree a summary is faithful, while missing details that are obvious errors only once pointed out. We therefore introduce a new dataset, StorySumm, comprising LLM summaries of short stories with localized faithfulness labels and error explanations. This benchmark is for evaluation methods, testing whether a given method can detect challenging inconsistencies. Using this dataset, we first show that any one human annotation protocol is likely to miss inconsistencies, and we advocate for pursuing a range of methods when establishing ground truth for a summarization dataset. We finally test recent automatic metrics and find that none of them achieve more than 70% balanced accuracy on this task, demonstrating that it is a challenging benchmark for future work in faithfulness evaluation.
MASIVE: Open-Ended Affective State Identification in English and Spanish
Nicholas Deas Columbia University, Elsbeth Turcan Columbia University, Ivan Ernesto Perez Mejia Columbia University, Kathleen McKeown Columbia University
Abstract:
In the field of emotion analysis, much NLP research focuses on identifying a limited number of discrete emotion categories, often applied across languages. These basic sets, however, are rarely designed with textual data in mind, and culture, language, and dialect can influence how particular emotions are interpreted. In this work, we broaden our scope to a practically unbounded set of affective states, which includes any terms that humans use to describe their experiences of feeling. We collect and publish MASIVE, a dataset of Reddit posts in English and Spanish containing over 1,000 unique affective states each. We then define the new problem of affective state identification for language generation models framed as a masked span prediction task. On this task, we find that smaller finetuned multilingual models outperform much larger LLMs, even on region-specific Spanish affective states. Additionally, we show that pretraining on MASIVE improves model performance on existing emotion benchmarks. Finally, through machine translation experiments, we find that native speaker-written data is vital to good performance on this task.
EmoKnob: Enhance Voice Cloning with Fine-Grained Emotion Control
Haozhe Chen Columbia University, Run Chen Columbia University, Julia Hirschberg Columbia University
Abstract:
While recent advances in Text-to-Speech (TTS) technology produce natural and expressive speech, they lack the option for users to select emotion and control intensity. We propose EmoKnob, a framework that allows fine-grained emotion control in speech synthesis with few-shot demonstrative samples of arbitrary emotion. Our framework leverages the expressive speaker representation space made possible by recent advances in foundation voice cloning models. Based on the few-shot capability of our emotion control framework, we propose two methods to apply emotion control on emotions described by open-ended text, enabling an intuitive interface for controlling a diverse array of nuanced emotions. To facilitate a more systematic emotional speech synthesis field, we introduce a set of evaluation metrics designed to rigorously assess the faithfulness and recognizability of emotion control frameworks. Through objective and subjective evaluations, we show that our emotion control framework effectively embeds emotions into speech and surpasses emotion expressiveness of commercial TTS services.
Enhancing Pre-Trained Generative Language Models with Question Attended Span Extraction on Machine Reading Comprehension
Lin Ai Columbia University, Zheng Hui Columbia University, Zizhou Liu Columbia University, Julia Hirschberg Columbia University
Abstract:
Machine Reading Comprehension (MRC) poses a significant challenge in the field of Natural Language Processing (NLP). While mainstream MRC methods predominantly leverage extractive strategies using encoder-only models such as BERT, generative approaches face the issue of out-of-control generation – a critical problem where answers generated are often incorrect, irrelevant, or unfaithful to the source text. To address these limitations in generative models for extractive MRC, we introduce the Question-Attended Span Extraction (QASE) module. Integrated during the finetuning phase of pre-trained generative language models (PLMs), QASE significantly enhances their performance, allowing them to surpass the extractive capabilities of advanced Large Language Models (LLMs) such as GPT-4 in few-shot settings. Notably, these gains in performance do not come with an increase in computational demands. The efficacy of the QASE module has been rigorously tested across various datasets, consistently achieving or even surpassing state-of-the-art (SOTA) results, thereby bridging the gap between generative and extractive models in extractive MRC tasks. Our code is available at this GitHub repository.
Defending Against Social Engineering Attacks in the Age of LLMs
Lin Ai Columbia University, Tharindu Sandaruwan Kumarage Arizona State University, Amrita Bhattacharjee Arizona State University, Zizhou Liu Columbia University, Zheng Hui Columbia University, Michael S. Davinroy Aptima, Inc., James Cook Aptima, Inc., Laura Cassani Aptima, Inc., Kirill Trapeznikov STR, Matthias Kirchner Kitware, Inc., Arslan Basharat Kirchner Kitware, Inc., Anthony Hoogs Kirchner Kitware, Inc., Joshua Garland Arizona State University, Huan Liu Arizona State University, Julia Hirschberg Columbia University
Abstract:
The proliferation of Large Language Models (LLMs) poses challenges in detecting and mitigating digital deception, as these models can emulate human conversational patterns and facilitate chat-based social engineering (CSE) attacks. This study investigates the dual capabilities of LLMs as both facilitators and defenders against CSE threats. We develop a novel dataset, SEConvo, simulating CSE scenarios in academic and recruitment contexts, and designed to examine how LLMs can be exploited in these situations. Our findings reveal that, while off-the-shelf LLMs generate high-quality CSE content, their detection capabilities are suboptimal, leading to increased operational costs for defense. In response, we propose ConvoSentinel, a modular defense pipeline that improves detection at both the message and the conversation levels, offering enhanced adaptability and cost-effectiveness. The retrievalaugmented module in ConvoSentinel identifies malicious intent by comparing messages to a database of similar conversations, enhancing CSE detection at all stages. Our study highlights the need for advanced strategies to leverage LLMs in cybersecurity. Our code and data are available at this GitHub repository.
LIONs: An Empirically Optimized Approach to Align Language Models
Xiao Yu Columbia University, Qingyang Wu Columbia University, Yu Li Columbia University, Zhou Yu Columbia University
Abstract:
Alignment is a crucial step to enhance the instruction-following and conversational abilities of language models. Despite many recent works proposing new algorithms, datasets, and training pipelines, there is a lack of comprehensive studies measuring the impact of various design choices throughout the whole training process. We first conduct a rigorous analysis over a three-stage training pipeline consisting of supervised fine-tuning, offline preference learning, and online preference learning. We have found that using techniques like sequence packing, loss masking in SFT, increasing the preference dataset size in DPO, and online DPO training can significantly improve the performance of language models. We then train from Gemma-2b-base and LLama-3-8b-base, and find that our best models exceed the performance of the official instruct models tuned with closed-source data and algorithms. Our code and models can be found at https://github.com/Columbia-NLP-Lab/LionAlignment.
DECOR: Improving Coherence in L2 English Writing with a Novel Benchmark for Incoherence Detection, Reasoning, and Rewriting
Xuanming Zhang Columbia University, Anthony Diaz University of California, Davis, Zixun Chen Columbia University, Qingyang Wu Columbia University, Kun Qian Columbia University, Erik Voss Columbia University, Zhou Yu Columbia University
Abstract:
Coherence in writing, an aspect that L2 English learners often struggle with, is crucial in assessing L2 English writing. Existing automated writing evaluation systems primarily use basic surface linguistic features to detect coherence in writing. However, little effort has been made to correct the detected incoherence, which could significantly benefit L2 language learners seeking to improve their writing. To bridge this gap, we introduce DECOR, a novel benchmark that includes expert annotations for detecting incoherence in L2 English writing, identifying the underlying reasons, and rewriting the incoherent sentences. To our knowledge, DECOR is the first coherence assessment dataset specifically designed for improving L2 English writing, featuring pairs of original incoherent sentences alongside their expert-rewritten counterparts. Additionally, we fine-tuned models to automatically detect and rewrite incoherence in student essays. We find that incorporating specific reasons for incoherence during fine-tuning consistently improves the quality of the rewrites, achieving a level that is favored in both automatic and human evaluations.
ACE: A LLM-based Negotiation Coaching System
Ryan Shea Columbia University, Aymen Kallala Columbia University, Xin Lucy Liu Columbia University, Michael W. Morris Columbia University, Zhou Yu Columbia University
Abstract:
The growing prominence of LLMs has led to an increase in the development of AI tutoring systems. These systems are crucial in providing underrepresented populations with improved access to valuable education. One important area of education that is unavailable to many learners is strategic bargaining related to negotiation. To address this, we develop a LLM-based Assistant for Coaching nEgotiation (ACE). ACE not only serves as a negotiation partner for users but also provides them with targeted feedback for improvement. To build our system, we collect a dataset of negotiation transcripts between MBA students. These transcripts come from trained negotiators and emulate realistic bargaining scenarios. We use the dataset, along with expert consultations, to design an annotation scheme for detecting negotiation mistakes. ACE employs this scheme to identify mistakes and provide targeted feedback to users. To test the effectiveness of ACE-generated feedback, we conducted a user experiment with two consecutive trials of negotiation and found that it improves negotiation performances significantly compared to a system that doesn’t provide feedback and one which uses an alternative method of providing feedback.
EDEN: Empathetic Dialogues for English Learning
Siyan Li Columbia University, Teresa Shao Columbia University, Zhou Yu Columbia University, Julia Hirschberg Columbia University
Abstract:
Dialogue systems have been used as conversation partners in English learning, but few have studied whether these systems improve learning outcomes. Student passion and perseverance, or grit, has been associated with language learning success. Recent work establishes that as students perceive their English teachers to be more supportive, their grit improves. Hypothesizing that the same pattern applies to English-teaching chatbots, we create EDEN, a robust open-domain chatbot for spoken conversation practice that provides empathetic feedback. To construct EDEN, we first train a specialized spoken utterance grammar correction model and a high-quality social chit-chat conversation model. We then conduct a preliminary user study with a variety of strategies for empathetic feedback. Our experiment suggests that using adaptive empathetic feedback leads to higher *perceived affective support*. Furthermore, elements of perceived affective support positively correlate with student grit.
A Fairness-Driven Method for Learning Human-Compatible Negotiation Strategies
Ryan Shea Columbia University, Zhou Yu Columbia University
Abstract:
Despite recent advancements in AI and NLP, negotiation remains a difficult domain for AI agents. Traditional game theoretic approaches that have worked well for two-player zero-sum games struggle in the context of negotiation due to their inability to learn human-compatible strategies. On the other hand, approaches that only use human data tend to be domain-specific and lack the theoretical guarantees provided by strategies grounded in game theory. Motivated by the notion of fairness as a criterion for optimality in general sum games, we propose a negotiation framework called FDHC which incorporates fairness into both the reward design and search to learn human-compatible negotiation strategies. Our method includes a novel, RL+search technique called LGM-Zero which leverages a pre-trained language model to retrieve human-compatible offers from large action spaces. Our results show that our method is able to achieve more egalitarian negotiation outcomes and improve negotiation quality.
TinyStyler: Efficient Few-Shot Text Style Transfer with Authorship Embeddings
Zachary Horvitz Columbia University, Ajay Patel University of Pennsylvania, Kanishk Singh Columbia University, Chris Callison-Burch University of Pennsylvania, Kathleen McKeown Columbia University, Zhou Yu Columbia University
Abstract
The goal of text style transfer is to transform the style of texts while preserving their original meaning, often with only a few examples of the target style. Existing style transfer methods generally rely on the few-shot capabilities of large language models or on complex controllable text generation approaches that are inefficient and underperform on fluency metrics. We introduce TinyStyler, a lightweight but effective approach, which leverages a small language model (800M params) and pre-trained authorship embeddings to perform efficient, few-shot text style transfer. We evaluate on the challenging task of authorship style transfer and find TinyStyler outperforms strong approaches such as GPT-4. We also evaluate TinyStyler’s ability to perform text attribute style transfer (formal ↔ informal) with automatic and human evaluations and find that the approach outperforms recent controllable text generation methods.
CS researchers presented their work at the 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024).
Parallel Structures in Pre-training Data Yield In-Context Learning
Yanda Chen Columbia University, Chen Zhao New York University, Zhou Yu Columbia University, Kathleen McKeown Columbia University, He He New York University
Abstract:
Pre-trained language models (LMs) are capable of in-context learning (ICL): they can adapt to a task with only a few examples given in the prompt without any parameter update. However, it is unclear where this capability comes from as there is a stark distribution shift between pre-training text and ICL prompts. In this work, we study what patterns of the pretraining data contribute to ICL. We find that LMs’ ICL ability depends on parallel structures in the pre-training data—pairs of phrases following similar templates in the same context window. Specifically, we detect parallel structures by checking whether training on one phrase improves prediction of the other, and conduct ablation experiments to study their effect on ICL. We show that removing parallel structures in the pre-training data reduces LMs’ ICL accuracy by 51% (vs 2% from random ablation). This drop persists even when excluding common patterns such as n-gram repetitions and long-range dependency, showing the diversity and generality of parallel structures. A closer look at the detected parallel structures indicates that they cover diverse linguistic tasks and span long distances in the data.
Getting Serious about Humor: Crafting Humor Datasets with Unfunny Large Language Models
Zachary Horvitz Columbia University, Jingru Chen Columbia University, Rahul Aditya Columbia University, Harshvardhan Srivastava Columbia University, Robert West EPFL, Zhou Yu Columbia University, Kathleen McKeown Columbia University
Abstract:
Humor is a fundamental facet of human cognition and interaction. Yet, despite recent advances in natural language processing, humor detection remains a challenging task that is complicated by the scarcity of datasets that pair humorous texts with similar non-humorous counterparts. We investigate whether large language models (LLMs) can generate synthetic data for humor detection via editing texts. We benchmark LLMs on an existing human dataset and show that current LLMs display an impressive ability to “unfun” jokes, as judged by humans and as measured on the downstream task of humor detection. We extend our approach to a code-mixed English-Hindi humor dataset where we find that GPT-4’s synthetic data is highly rated by bilingual annotators and provides challenging adversarial examples for humor classifiers.
ICLEF: In-Context Learning with Expert Feedback for Explainable Style Transfer
Arkadiy Saakyan Columbia University and Smaranda Muresan Columbia University
Abstract:
While state-of-the-art large language models (LLMs) can excel at adapting text from one style to another, current work does not address the explainability of style transfer models. Recent work has explored generating textual explanations from larger teacher models and distilling them into smaller student models. One challenge with such approach is that LLM outputs may contain errors that require expertise to correct, but gathering and incorporating expert feedback is difficult due to cost and availability. To address this challenge, we propose ICLEF, a novel human-AI collaboration approach to model distillation that incorporates scarce expert human feedback by combining in-context learning and model self-critique. We show that our method leads to generation of high-quality synthetic explainable style transfer datasets for formality (E-GYAFC) and subjective bias (EWNC). Via automatic and human evaluation, we show that specialized student models finetuned on our datasets outperform generalist teacher models on the explainable style transfer task in one-shot settings, and perform competitively compared to few-shot teacher models, highlighting the quality of the data and the role of expert feedback. In an extrinsic task of authorship attribution, we show that explanations generated by smaller models fine-tuned on E-GYAFC are more predictive of authorship than explanations generated by few-shot teacher models.
ProLex: A Benchmark for Language Proficiency-oriented Lexical Substitution
Xuanming Zhang Columbia University, Zixun Chen Columbia University, and Zhou Yu Columbia University
Abstract:
Lexical Substitution discovers appropriate substitutes for a given target word in a context sentence. However, the task fails to consider substitutes that are of equal or higher proficiency than the target, an aspect that could be beneficial for language learners looking to improve their writing. To bridge this gap, we propose a new task — language proficiencyoriented lexical substitution. We also introduce ProLex, a novel benchmark designed to assess systems’ ability to generate not only appropriate substitutes but also substitutes that demonstrate better language proficiency. Besides the benchmark, we propose models that can automatically perform the new task. We show that our best model, a Llama2-13B model fine-tuned with task-specific synthetic data, outperforms ChatGPT by an average of 3.2% in F-score and achieves comparable results with GPT-4 on ProLex.
LOCALRQA: From Generating Data to Locally Training, Testing, and Deploying Retrieval-Augmented QA Systems
Xiao Yu Columbia University, Yunan Lu Columbia University, Zhou Yu Columbia University
Abstract:
Retrieval-augmented question-answering systems combine retrieval techniques with large language models to provide answers that are more accurate and informative. Many existing toolkits allow users to quickly build such systems using off-the-shelf models, but they fall short in supporting researchers and developers to customize the model training, testing, and deployment process. We propose LOCALRQA1 , an open-source toolkit that features a wide selection of model training algorithms, evaluation methods, and deployment tools curated from the latest research. As a showcase, we build QA systems using online documentation obtained from Databricks and Faire’s websites. We find 7B-models trained and deployed using LOCALRQA reach a similar performance compared to using OpenAI’s text-ada-002 and GPT-4-turbo.
Large Language Models are Few-Shot Training Example Generators: A Case Study in Fallacy Recognition
Tariq Alhindi Mohamed bin Zayed University of Artificial Intelligence, Smaranda Muresan Columbia University, Preslav Nakov Mohamed bin Zayed University of Artificial Intelligence
Abstract:
Recognizing fallacies is crucial for ensuring the quality and validity of arguments across various domains. However, computational fallacy recognition faces challenges due to the diverse genres, domains, and types of fallacies found in datasets. This leads to a highly multiclass, and even multi-label, setup with substantial class imbalance. In this study, we aim to enhance existing models for fallacy recognition by incorporating additional context and by leveraging large language models to generate synthetic data, thus increasing the representation of the infrequent classes. We experiment with GPT-3.5 to generate synthetic examples and we examine the impact of prompt settings for this. Moreover, we explore zero-shot and few-shot scenarios to evaluate the effectiveness of using the generated examples for training smaller models within a unified fallacy recognition framework. Furthermore, we analyze the overlap between the synthetic data and existing fallacy datasets. Finally, we investigate the usefulness of providing supplementary context for detecting fallacy types that need such context, e.g., diversion fallacies. Our evaluation results demonstrate consistent improvements across fallacy types, datasets, and generators. The code and the synthetic datasets are all publicly available.
Papers from CS researchers were accepted to the 41st International Conference on Machine Learning (ICML 2024). They join the machine learning research community in Vienna, Austria, on July 21 – 27, 2024. ICML brings together the brightest minds in the field to share their latest findings, foster collaborations, and inspire new directions in machine learning.
The links to the papers and the abstracts are below:
SelfIE: Self-Interpretation of Large Language Model Embeddings
Haozhe Chen Columbia University, Carl Vondrick Columbia University, Chengzhi Mao Columbia University
Abstract:
How do large language models (LLMs) obtain their answers? The ability to explain and control an LLM’s reasoning process is key for reliability, transparency, and future model developments. We propose SelfIE (Self-Interpretation of Embeddings), a framework that enables LLMs to interpret their own embeddings in natural language by leveraging their ability to respond to inquiries about a given passage. Capable of interpreting open-world concepts in the hidden embeddings, SelfIE reveals LLM internal reasoning in cases such as making ethical decisions, internalizing prompt injection, and recalling harmful knowledge. SelfIE’s text descriptions on hidden embeddings open avenues to control LLM reasoning. We propose Supervised Control, which allows editing open-ended concepts while only requiring gradient computation of individual layer. We extend RLHF to hidden embeddings and propose Reinforcement Control that erases harmful knowledge in LLM without supervision targets.
Counterfactual Image Editing
Yushu Pan Columbia University, Elias Bareinboim Columbia University
Abstract:
Counterfactual image editing is a challenging task within generative AI. The current literature on the topic focuses primarily on changing individual features while being silent about the causal relationships between features, which are present in the real world. In this paper, we first formalize this task through causal language, modeling the causal relationships between latent generative factors and images through a special type of causal model called augmented structural causal models (ASCMs). Second, we show two fundamental impossibility results: (1) counterfactual editing is impossible from i.i.d. image samples and their corresponding labels alone; (2) also, even when the causal relationships between latent generative factors and images are available, no guarantees regarding the output of the generative model can be provided. Third, we propose a relaxation over this hard problem aiming to approximate the non-identifiable target counterfactual distributions while still preserving features the users care about and that are causally consistent with the true generative model, which we call ctf-consistent estimators. Finally, we develop an efficient algorithm to generate counterfactual image samples leveraging neural causal models.
Exploiting Code Symmetries for Learning Program Semantics
Kexin Pei Columbia University, Weichen Li Columbia University, Qirui Jin University of Michigan, Shuyang Liu Huazhong University of Science and Technology, Scott Geng Univerisity of Washington, Lorenzo Cavallaro University College London, Junfeng Yang Columbia University, Suman Jana Columbia University
Abstract:
This paper tackles the challenge of teaching code semantics to Large Language Models (LLMs) for program analysis by incorporating code symmetries into the model architecture. We introduce a group-theoretic framework that defines code symmetries as semantics-preserving transformations, where forming a code symmetry group enables precise and efficient reasoning of code semantics. Our solution, SymC, develops a novel variant of self-attention that is provably equivariant to code symmetries from the permutation group defined over the program dependence graph. SymC obtains superior performance on five program analysis tasks, outperforming state-of-the-art code models, including GPT-4, without any pre-training. Our results suggest that code LLMs that encode the code structural prior via the code symmetry group generalize better and faster.
MGit: A Model Versioning and Management System
Wei Hao Columbia University, Daniel Mendoza Stanford University, Rafael Mendes Microsoft Research, Deepak Narayanan NVIDIA, Amar Phanishayee Columbia University, Asaf Cidon Columbia University, Junfeng Yang Columbia University
Abstract:
New ML models are often derived from existing ones (e.g., through fine-tuning, quantization or distillation), forming an ecosystem where models are *related* to each other and can share structure or even parameter values. Managing such a large and evolving ecosystem of model derivatives is challenging. For instance, the overhead of storing all such models is high, and models may inherit bugs from related models, complicating error attribution and debugging. In this paper, we propose a model versioning and management system called MGit that makes it easier to store, test, update, and collaborate on related models. MGit introduces a lineage graph that records the relationships between models, optimizations to efficiently store model parameters, and abstractions over this lineage graph that facilitate model testing, updating and collaboration. We find that MGit works well in practice: MGit is able to reduce model storage footprint by up to 7x. Additionally, in a user study with 20 ML practitioners, users complete a model updating task 3x faster on average with MGit.
Position: TrustLLM: Trustworthiness in Large Language Models
Yue Huang Lehigh University, Lichao Sun Lehigh University, Haoran Wang Illinois Institute of Technology, Siyuan Wu CISPA, Qihui Zhang CISPA, Yuan Li University of Cambridge, Chujie Gao CISPA, Yixin Huang Institut Polytechnique de Paris, Wenhan Lyu William & Mary, Yixuan Zhang William & Mary, Xiner Li Texas A&M University, Hanchi Sun Lehigh University, Zhengliang Liu University of Georgia, Yixin Liu Lehigh University, Yijue Wang Samsung Research America, Zhikun Zhang Stanford University, Bertie Vidgen MLCommons, Bhavya Kailkhura Lawrence Livermore National Laboratory, Caiming Xiong Salesforce Research, Chaowei Xiao University of Wisconsin, Madison, Chunyuan Li Microsoft Research, Eric Xing Carnegie Mellon University, Furong Huang University of Maryland, Hao Liu University of California, Berkeley, Heng Ji University of Illinois Urbana-Champaign, Hongyi Wang Rutgers University, Huan Zhang University of Illinois Urbana-Champaign, Huaxiu Yao UNC Chapel Hill, Manolis Kellis Massachusetts Institute of Technology, Marinka Zitnik Harvard University, Meng Jiang University of Notre Dame, Mohit Bansal UNC Chapel Hill, James Zou Stanford University, Jian Pei Duke University, Jian Liu University of Tennessee, Knoxville, Jianfeng Gao Microsoft Research, Jiawei Han University of Illinois Urbana-Champaign, Jieyu Zhao University of Southern California, Jiliang Tang Michigan State University, Jindong Wang Microsoft Research Asia, Joaquin Vanschoren Eindhoven University of Technology, John Mitchell Drexel University, Kai Shu Illinois Institute of Technology, Kaidi Xu Drexel University, Kai-Wei Chang University of California, Los Angeles, Lifang He Lehigh University, Lifu Huang Virginia Tech, Michael Backes CISPA, Neil Gong Duke University, Philip Yu University of Illinois Chicago, Pin-Yu Chen IBM Research, Quanquan Gu University of California, Los Angeles, Ran Xu Salesforce Research, Rey Ying Yale University, Shuiwang Ji Texas A&M University, Suman Jana Columbia UniversityI, Tianlong Chen UNC Chapel Hill, Tianming Liu University of Georgia, Tianyi Zhou University of Maryland, William Wang University of California, Santa Barbara, Xiang Li Massachusetts General Hospital, Xiangliang Zhang University of Notre Dame, Xiao Wang Northwestern University, Xing Xie Microsoft Research Asia, Xun Chen Samsung Research America, Xuyu Wang Florida International University, Yan Liu University of Southern California, Yanfang Ye University of Notre Dame, Yinzhi Cao Johns Hopkins University, Yong Chen University of Pennsylvania, Yue Zhao University of Southern California
Abstract:
Large language models (LLMs) have gained considerable attention for their excellent natural language processing capabilities. Nonetheless, these LLMs present many challenges, particularly in the realm of trustworthiness. This paper introduces TrustLLM, a comprehensive study of trustworthiness in LLMs, including principles for different dimensions of trustworthiness, established benchmark, evaluation, and analysis of trustworthiness for mainstream LLMs, and discussion of open challenges and future directions. Specifically, we first propose a set of principles for trustworthy LLMs that span eight different dimensions. Based on these principles, we further establish a benchmark across six dimensions including truthfulness, safety, fairness, robustness, privacy, and machine ethics. We then present a study evaluating 16 mainstream LLMs in TrustLLM, consisting of over 30 datasets. Our findings firstly show that in general trustworthiness and capability (i.e., functional effectiveness) are positively related. Secondly, our observations reveal that proprietary LLMs generally outperform most open-source counterparts in terms of trustworthiness, raising concerns about the potential risks of widely accessible open-source LLMs. However, a few open-source LLMs come very close to proprietary ones, suggesting that open-source models can achieve high levels of trustworthiness without additional mechanisms like moderator, offering valuable insights for developers in this field. Thirdly, it is important to note that some LLMs may be overly calibrated towards exhibiting trustworthiness, to the extent that they compromise their utility by mistakenly treating benign prompts as harmful and consequently not responding. Besides these observations, we’ve uncovered key insights into the multifaceted trustworthiness in LLMs. We emphasize the importance of ensuring transparency not only in the models themselves but also in the technologies that underpin trustworthiness. We advocate that the establishment of an AI alliance between industry, academia, the open-source community to foster collaboration is imperative to advance the trustworthiness of LLMs.
Add to GMail