ACL 2024 论文列表

In this study, we explore the application of Large Language Models (LLMs) in Jubensha, a Chinese detective role-playing game and a novel area in Artificial Intelligence (AI) driven gaming. We introduce the first dataset specifically for Jubensha, including character scripts and game rules, to foster AI agent development in this complex narrative environment. Our work also presents a unique multi-agent interaction framework using LLMs, allowing AI agents to autonomously engage in Jubensha games. To evaluate the gaming performance of these AI agents, we developed novel methods measuring their mastery of case information and reasoning skills. Furthermore, we incorporated the latest advancements in prompting engineering to enhance the agents’ performance in information gathering, murderer identification, and logical reasoning. The experimental results validate the effectiveness of our proposed methods. This work aims to offer a novel perspective on understanding LLM capabilities and establish a new benchmark for evaluating large language model-based agents.

Large Language Models (LLMs) have demonstrated exceptional coding capability. However, as another critical component of programming proficiency, the debugging capability of LLMs remains relatively unexplored. Previous evaluations of LLMs’ debugging ability are significantly limited by the risk of data leakage, the scale of the dataset, and the variety of tested bugs. To overcome these deficiencies, we introduce ‘DebugBench’, an LLM debugging benchmark consisting of 4,253 instances. It covers four major bug categories and 18 minor types in C++, Java, and Python. To construct DebugBench, we collect code snippets from the LeetCode community, implant bugs into source data with GPT-4, and assure rigorous quality checks. We evaluate two commercial and four open-source models in a zero-shot scenario. We find that (1) while closed-source models exhibit inferior debugging performance compared to humans, open-source models relatively lower pass rate scores; (2) the complexity of debugging notably fluctuates depending on the bug category; (3) incorporating runtime feedback has a clear impact on debugging performance which is not always helpful. As an extension, we also compare LLM debugging and code generation, revealing a strong correlation between them for closed-source models. These findings will benefit the development of LLMs in debugging.

Large language models (LLMs) are leading significant progress in code generation. Beyond one-pass code generation, recent works further integrate unit tests and program verifiers into LLMs to iteratively refine the generated programs. However, these works consider the generated programs as an indivisible entity, which falls short for LLMs in debugging the programs, especially when the programs contain complex logic flows and data operations. In contrast, when human developers debug programs, they typically set breakpoints and selectively examine runtime execution information. The execution flow and the intermediate variables play a crucial role in the debugging process, yet they are underutilized in the existing literature on code generation. In this study, we introduce Large Language Model Debugger (LDB), a novel debugging framework that enables LLMs to refine their generated programs with the runtime execution information. Specifically, LDB segments the programs into basic blocks and tracks the values of intermediate variables after each block throughout the runtime execution. This allows LLMs to concentrate on simpler code units within the overall execution flow, verify their correctness against the task description block by block, and efficiently pinpoint any potential errors. Experiments demonstrate that LDB consistently enhances the baseline performance by up to 9.8% across the HumanEval, MBPP, and TransCoder benchmarks, archiving new state-of-the-art performance in code debugging for various LLM selections.

Due to biases inherently present in data for pre-training, current pre-trained Large Language Models (LLMs) also ubiquitously manifest the same phenomena. Since the bias influences the output from the LLMs across various tasks, the widespread deployment of the LLMs is hampered. We propose a simple method that utilizes structured knowledge to alleviate this issue, aiming to reduce the bias embedded within the LLMs and ensuring they have an encompassing perspective when used in applications. Experimental results indicated that our method has good debiasing ability when applied to existing both autoregressive and masked language models. Additionally, it could ensure that the performances of LLMs on downstream tasks remain uncompromised.Our method outperforms state-of-the-art (SOTA) baselines in the debiasing ability. Importantly, our method obviates the need for training from scratch, thus offering enhanced scalability and cost-effectiveness.

In-context learning(ICL) has gained considerable attention due to its data efficiency and task adaptability. Unfortunately, ICL suffers from the demonstration bias, i.e., its performance and robustness are severely affected by the selection and ordering of demonstrations. In this paper, we identify that such demonstration bias may primarily stem from the semantic ambiguity induced by demonstrations, i.e., a demonstration may indicate multiple input-to-label mappings and its mapping can be interpreted differently in different contexts by LLMs. Such semantic ambiguity disrupts task comprehension during ICL and results in performance fluctuations. To resolve the semantic ambiguity problem, this paper further proposes two de-biasing strategies to mitigate demonstration bias in in-context learning. Experiments on six datasets show that our methods can effectively alleviate demonstration bias and significantly improve task performance.

How can we construct an automated debate judge to evaluate an extensive, vibrant, multi-turn debate? This task is challenging, as judging a debate involves grappling with lengthy texts, intricate argument relationships, and multi-dimensional assessments.At the same time, current research mainly focuses on short dialogues, rarely touching upon the evaluation of an entire debate.In this paper, by leveraging Large Language Models (LLMs), we propose Debatrix, which makes the analysis and assessment of multi-turn debates more aligned with majority preferences. Specifically, Debatrix features a vertical, iterative chronological analysis and a horizontal, multi-dimensional evaluation collaboration.To align with real-world debate scenarios, we introduced the PanelBench benchmark, comparing our system’s performance to actual debate outcomes.The findings indicate a notable enhancement over directly using LLMs for debate evaluation.Source code and benchmark data are available at https://github.com/ljcleo/debatrix.

Effective interlocutors account for the uncertain goals, beliefs, and emotions of others. But even the best human conversationalist cannot perfectly anticipate the trajectory of a dialogue. How well can language models represent inherent uncertainty in conversations? We propose FortUne Dial, an expansion of the long-standing “conversation forecasting” task: instead of just accuracy, evaluation is conducted with uncertainty-aware metrics, effectively enabling abstention on individual instances. We study two ways in which language models potentially represent outcome uncertainty (internally, using scores and directly, using tokens) and propose fine-tuning strategies to improve calibration of both representations. Experiments on eight difficult negotiation corpora demonstrate that our proposed fine-tuning strategies (a traditional supervision strategy and an off-policy reinforcement learning strategy) can calibrate smaller open-source models to compete with pre-trained models 10x their size.

We present a novel human annotated dataset for evaluating the ability for visual-language models to generate both short and long descriptions for real-world video clips, termed DeVAn (Dense Video Annotation). The dataset contains 8.5K YouTube video clips of 20-60 seconds in duration and covers a wide range of topics and interests. Each video clip is independently annotated by 5 human annotators, producing both captions (1 sentence) and summaries (3-10 sentences). Given any video selected from the dataset and its corresponding ASR information, we evaluate visual-language models on either caption or summary generation that is grounded in both the visual and auditory content of the video. Additionally, models are also evaluated on caption- and summary-based retrieval tasks, where the summary-based retrieval task requires the identification of a target video given excerpts of a given summary. Given the novel nature of the paragraph-length video summarization task, we compared different existing evaluation metrics and their alignment with human preferences and found that model-based evaluation metrics provide more semantically-oriented and human-aligned evaluation. Finally, we benchmarked a wide range of current video-language models on DeVAn, and we aim for DeVAn to serve as a useful evaluation set in the age of large language models and complex multi-modal tasks. Code is available at https://github.com/TK-21st/DeVAn.

Large language models (LLMs) often require task-relevant knowledge to augment their internal knowledge through prompts. However, simply injecting external knowledge into prompts does not guarantee that LLMs can identify and use relevant information in the prompts to conduct chain-of-thought reasoning, especially when the LLM’s internal knowledge is derived from biased information on the pretraining data. In this paper, we propose a novel causal view to formally explain the internal knowledge bias of LLMs via a Structural Causal Model (SCM). We review the chain-of-thought (CoT) prompting from a causal perspective and discover that the biased information from pretrained models can impair LLMs’ reasoning abilities. When the CoT reasoning paths are misled by irrelevant information from prompts and are logically incorrect, simply editing factual information is insufficient to reach the correct answer. To estimate the confounding effect on CoT reasoning in LLMs, we use external knowledge as an instrumental variable. We further introduce CoT as a mediator to conduct front-door adjustment and generate logically correct CoTs where the spurious correlation between LLMs’ pretrained knowledge and task queries is reduced. With extensive experiments, we validate that our approach enables more accurate CoT reasoning and enhances LLM generation on knowledge-intensive tasks.

Recent years have witnessed the deployment of code language models (LMs) in various code intelligence tasks such as code completion. Yet, it is challenging for pre-trained LMs to generate correct completions in private repositories. Previous studies retrieve cross-file context based on import relations or text similarity, which is insufficiently relevant to completion targets. In this paper, we propose a dataflow-guided retrieval augmentation approach, called DraCo, for repository-level code completion. DraCo parses a private repository into code entities and establishes their relations through an extended dataflow analysis, forming a repo-specific context graph. Whenever triggering code completion, DraCo precisely retrieves relevant background knowledge from the repo-specific context graph and generates well-formed prompts to query code LMs. Furthermore, we construct a large Python dataset, ReccEval, with more diverse completion targets. Our experiments demonstrate the superior accuracy and applicable efficiency of DraCo, improving code exact match by 3.43% and identifier F1-score by 3.27% on average compared to the state-of-the-art approach.

Large language models (LLMs) have become a dominant and important tool for NLP researchers in a wide range of tasks. Today, many researchers use LLMs in synthetic data generation, task evaluation, fine-tuning, distillation, and other model-in-the-loop research workflows. However, challenges arise when using these models that stem from their scale, their closed source nature, and the lack of standardized tooling for these new and emerging workflows. The rapid rise to prominence of these models and these unique challenges has had immediate adverse impacts on open science and on the reproducibility of work that uses them. In this ACL 2024 theme track paper, we introduce DataDreamer, an open source Python library that allows researchers to write simple code to implement powerful LLM workflows. DataDreamer also helps researchers adhere to best practices that we propose to encourage open science and reproducibility. The library and documentation are available at: https://github.com/datadreamer-dev/DataDreamer.

Human-like social bias of pre-trained language models (PLMs) on downstream tasks have attracted increasing attention. The potential flaws in the training data are the main factor that causes unfairness in PLMs. Existing data-centric debiasing strategies mainly leverage explicit bias words (defined as sensitive attribute words specific to demographic groups) for counterfactual data augmentation to balance the training data. However, they lack consideration of implicit bias words potentially associated with explicit bias words in complex distribution data, which indirectly harms the fairness of PLMs. To this end, we propose a **Data**-Centric **Debias**ing method (named Data-Debias), which uses an explainability method to search for implicit bias words to assist in debiasing PLMs. Specifically, we compute the feature attributions of all tokens using the Integrated Gradients method, and then treat the tokens that have a large impact on the model’s decision as implicit bias words. To make the search results more precise, we iteratively train a biased model to amplify the bias with each iteration. Finally, we use the implicit bias words searched in the last iteration to assist in debiasing PLMs. Extensive experimental results on multiple PLMs debiasing on three different classification tasks demonstrate that Data-Debias achieves state-of-the-art debiasing performance and strong generalization while maintaining predictive abilities.

The rapid advancements of Large Language Models (LLMs) tightly associate with the expansion of the training data size. However, the unchecked ultra-large-scale training sets introduce a series of potential risks like data contamination, i.e. the benchmark data is used for training. In this work, we propose a holistic method named Polarized Augment Calibration (PAC) along with a new to-be-released dataset to detect the contaminated data and diminish the contamination effect. PAC extends the popular MIA (Membership Inference Attack) — from machine learning community — by forming a more global target at detecting training data to Clarify invisible training data. As a pioneering work, PAC is very much plug-and-play that can be integrated with most (if not all) current white- and black-box LLMs. By extensive experiments, PAC outperforms existing methods by at least 4.5%, towards data contamination detection on more 4 dataset formats, with more than 10 base LLMs. Besides, our application in real-world scenarios highlights the prominent presence of contamination and related issues.

In the rapidly evolving field of large language models (LLMs), data augmentation (DA) has emerged as a pivotal technique for enhancing model performance by diversifying training examples without the need for additional data collection. This survey explores the transformative impact of LLMs on DA, particularly addressing the unique challenges and opportunities they present in the context of natural language processing (NLP) and beyond. From both data and learning perspectives, we examine various strategies that utilize LLMs for data augmentation, including a novel exploration of learning paradigms where LLM-generated data is used for diverse forms of further training. Additionally, this paper highlights the primary open challenges faced in this domain, ranging from controllable data augmentation to multi-modal data augmentation. This survey highlights a paradigm shift introduced by LLMs in DA, and aims to serve as a comprehensive guide for researchers and practitioners.

Recently, large language models (LLMs) enhanced by self-reflection have achieved promising performance on machine transla004 tion. The key idea is guiding LLMs to generate translation with human-like feedback. However, existing self-reflection methods lack effective feedback information, limiting the translation performance. To address this, we introduce a DUAL-REFLECT framework, leveraging the dual learning of translation tasks to provide effective feedback, thereby enhancing the models’ self-reflective abilities and improving translation performance. The application of this method across various translation tasks has proven its effectiveness in improving translation accuracy and eliminating ambiguities, especially in translation tasks with low-resource language pairs.

Dynamic retrieval augmented generation (RAG) paradigm actively decides when and what to retrieve during the text generation process of Large Language Models (LLMs).There are two key elements of this paradigm: identifying the optimal moment to activate the retrieval module (deciding when to retrieve) and crafting the appropriate query once retrieval is triggered (determining what to retrieve).However, current dynamic RAG methods fall short in both aspects. Firstly, the strategies for deciding when to retrieve often rely on static rules. Moreover, the strategies for deciding what to retrieve typically limit themselves to the LLM’s most recent sentence or the last few tokens, while the LLM’s information needs may span across the entire context.To overcome these limitations, we introduce a new framework, DRAGIN, i.e., Dynamic Retrieval Augmented Generation based on the Information Needs of LLMs. Our framework is specifically designed to make decisions on when and what to retrieve based on the LLM’s information needs during the text generation process.We evaluate DRAGIN along with existing methods comprehensively over 4 knowledge-intensive generation datasets. Experimental results show that DRAGIN achieves superior performance on all tasks, demonstrating the effectiveness of our method.

The success of large language models (LLM) benefits from large-scale model parameters and large amounts of pre-training data. However, the textual data for training LLM can not be confirmed to be legal because they are crawled from different web sites. For example, there are copyrighted articles, personal reviews and information in the pre-training data for LLM which are illegal. To address the above issue and develop legal LLM, we propose to detect the pre-training data from LLM in a pure black-box way because the existing LLM services only return the generated text. The previous most related works are the membership inference attack (MIA) on machine learning models to detect the training data from them. But the existing methods are based on analyzing the output probabilities of models which are unrealistic to LLM services. To tackle the problem, we firstly construct the benchmark datasets by collecting textual data from different domains as the seen and unseen pre-training data for LLMs. Then, we investigate a black-box framework named DPDLLM, with the only access to the generated texts from LLM for detecting textual data whether was used to train it. In the proposed framework, we exploit GPT-2 as the reference model to fit the textual data and feed the generated text from LLM into it to acquire sequence probabilities as the significant feature for detection. The experimental results on the benchmark datasets demonstrate that DPDLLM is effective on different popular LLMs and outperforms the existing methods.

Prompt recovery in large language models (LLMs) is crucial for understanding how LLMs work and addressing concerns regarding privacy, copyright, etc. The trend towards inference-only APIs complicates this task by restricting access to essential outputs for recovery. To tackle this challenge, we extract prompt-related information from limited outputs and identify a strong(negative) correlation between output probability-based uncertainty and the success of prompt recovery.This finding led to the development of Deliberative PrOmpt RecoverY (DORY), our novel approach that leverages uncertainty to recover prompts accurately. DORY involves reconstructing drafts from outputs, refining these with hints, and filtering out noise based on uncertainty. Our evaluation shows that DORY outperforms existing baselines across diverse LLMs and prompt benchmarks, improving performance by approximately 10.82% and establishing a new state-of-the-art record in prompt recovery tasks. Significantly, DORY operates using a single LLM without any external resources or model, offering a cost-effective, user-friendly prompt recovery solution.

The performance of the reward model (RM) is a critical factor in improving the effectiveness of the large language model (LLM) during alignment fine-tuning. There remain two challenges in RM training: 1) training the same RM using various categories of data may cause its generalization performance to suffer from multi-task disturbance, and 2) the human annotation consistency rate is generally only 60% to 75%, causing training data to contain a lot of noise. To tackle these two challenges, we introduced the idea of Mixture-of-Experts (MoE) into the field of RM for the first time. We propose the Double-Layer MoE RM (DMoERM). The outer layer MoE is a sparse model. After classifying an input into task categories, we route it to the corresponding inner layer task-specific model. The inner layer MoE is a dense model. We decompose the specific task into multiple capability dimensions and individually fine-tune a LoRA expert on each one. Their outputs are then synthesized by an MLP to compute the final rewards. To minimize costs, we call a public LLM API to obtain the capability preference labels. The validation on manually labeled datasets confirms that our model attains superior consistency with human preference and outstrips advanced generative approaches. Meanwhile, through BoN sampling and RL experiments, we demonstrate that our model outperforms state-of-the-art ensemble methods of RM and mitigates the overoptimization problem. Our code is available at: https://github.com/quanshr/DMoERM.