论文检索

Liyan Xu, Zhenlin Su, Mo Yu, Jin Xu, Jinho D. Choi, Jie Zhou, Fei Liu

Factual inconsistencies pose a significant hurdle for the faithful summarization by generative models. While a major direction to enhance inconsistency detection is to derive stronger Natural Language Inference (NLI) models, we propose an orthogonal aspect that underscores the importance of incorporating task-specific taxonomy into the inference. To this end, we consolidate key error types of inconsistent facts in summaries, and incorporate them to facilitate both the zero-shot and supervised paradigms of LLMs. Extensive experiments on ten datasets of five distinct domains suggest that, zero-shot LLM inference could benefit from the explicit solution space depicted by the error type taxonomy, and achieves state-of-the-art performance overall, surpassing specialized non-LLM baselines, as well as recent LLM baselines. We further distill models that fuse the taxonomy into parameters through our designed prompt completions and supervised training strategies, efficiently substituting state-of-the-art zero-shot inference with much larger LLMs.

Yue Jiang, Ziyu Guan, Jie Zhao, Wei Zhao, Jiaqi Yang

Legal question answering (LQA) aims to bridge the gap between the limited availability of legal professionals and the high demand for legal assistance. Traditional LQA approaches typically either select the optimal answers from an answer set or extract answers from law texts. However, they often struggle to provide relevant answers to complex, real-world questions due to the rigidity of predetermined answers. Although recent advancements in legal large language models have shown some potential in enhancing answer relevance, they fail to address the multiple user-specific circumstances, i.e., factual details in questions. To address these issues, we (1) construct the first publicly available legal community question-answering (LegalCQA) dataset; and (2) propose a Hierarchical Legal Knowledge Integration (H-LegalKI) framework. LegalCQA is collected from two widely used legal forums for developing user-centered LQA models. For H-LegalKI, we design a legal knowledge retriever that gathers comprehensive legal knowledge based on both entire questions and individual sentences. And an answer generation model is designed to understand question- and sentence-level factual details and integrate corresponding legal knowledge in a hierarchical way. Additionally, we design a de-redundancy module to remove redundant legal knowledge. Experiments on LegalCQA demonstrate the superiority of our framework over competitive baselines.

Shruti Singh, Shoaib Alam, Husain Malwat, Mayank Singh

The ever-increasing volume of paper submissions makes it difficult to stay informed about the latest state-of-the-art research. To address this challenge, we introduce LEGOBench, a benchmark for evaluating systems that generate scientific leaderboards. LEGOBench is curated from 22 years of preprint submission data on arXiv and more than 11k machine learning leaderboards on the PapersWithCode portal. We present a language model-based and four graph-based leaderboard generation task configuration. We evaluate popular encoder-only scientific language models as well as decoder-only large language models across these task configurations. State-of-the-art models showcase significant performance gaps in automatic leaderboard generation on LEGOBench. The code is available on GitHub and the dataset is hosted on OSF.

Daniel Gareev, Thomas Hofmann, Ezhilmathi Krishnasamy, Tiago Pimentel

Text generation, a component in applications such as dialogue systems, relies heavily on decoding algorithms that sample strings from a language model distribution. Traditional methods like top-k and top-\pi decoding locally normalise the model’s output, which can significantly distort the original distribution. In this paper, we investigate the effects of such distortions by introducing globally-normalised versions of these decoding methods. Further, we propose an independent Metropolis-Hastings (IMH) algorithm to approximate sampling from these globally-normalised distributions without explicitly computing them. Our empirical analyses compare the performance of local and global decoding across two algorithms (top-k and top-\pi) with various hyperparameters, using the Pythia language models. Results show that in most configuration, global decoding performs worse than the local decoding versions of the same algorithms, despite preserving the distribution’s integrity. Our results thus suggest that distortion might be an important feature of local decoding algorithms.

Yiting Ran, Xintao Wang, Rui Xu, Xinfeng Yuan, Jiaqing Liang, Yanghua Xiao, Deqing Yang

Role-playing agents (RPA) have been a popular application area for large language models (LLMs), attracting significant interest from both industry and academia. While existing RPAs well portray the characters’ knowledge and tones, they face challenges in capturing their minds, especially for small role-playing language models (RPLMs). In this paper, we propose to enhance RPLMs via personality-indicative data. Specifically, we leverage questions from psychological scales and distill advanced RPAs to generate dialogues that grasp the minds of characters. Experimental results validate that RPLMs trained with our dataset exhibit advanced role-playing capabilities for both general and personality-related evaluations.

Anwoy Chatterjee, H S V N S Kowndinya Renduchintala, Sumit Bhatia, Tanmoy Chakraborty

Despite their remarkable capabilities, Large Language Models (LLMs) are found to be surprisingly sensitive to minor variations in prompts, often generating significantly divergent outputs in response to minor variations in the prompts, such as spelling errors, alteration of wording or the prompt template. However, while assessing the quality of an LLM, the focus often tends to be solely on its performance on downstream tasks, while very little to no attention is paid to prompt sensitivity. To fill this gap, we propose POSIX – a novel PrOmpt Sensitivity IndeX as a reliable measure of prompt sensitivity, thereby offering a more comprehensive evaluation of LLM performance. The key idea behind POSIX is to capture the relative change in loglikelihood of a given response upon replacing the corresponding prompt with a different intent-preserving prompt. We provide thorough empirical evidence demonstrating the efficacy of POSIX in capturing prompt sensitivity and subsequently use it to measure and thereby compare prompt sensitivity of various open source LLMs. We find that merely increasing the parameter count or instruction tuning does not necessarily reduce prompt sensitivity whereas adding some few-shot exemplars, even just one, almost always leads to significant decrease in prompt sensitivity. We also find that alterations to prompt template lead to the highest sensitivity in the case of MCQ type tasks, whereas paraphrasing results in the highest sensitivity in open-ended generation tasks. The code for reproducing our results is open-sourced at https://github.com/kowndinya-renduchintala/POSIX.

Xiao Pu, Tianxing He, Xiaojun Wan

Prompt compression condenses contexts while maintaining their informativeness for different usage scenarios. It not only shortens the inference time and reduces computational costs during the usage of large language models, but also lowers expenses when using closed-source models. In a preliminary study, we discover that when instructing language models to compress prompts, different compression styles (e.g., extractive or abstractive) impact performance of compressed prompts on downstream tasks. Building on this insight, we propose Style-Compress, a lightweight framework that adapts a smaller language model to compress prompts for a larger model on a new task without additional training. Our approach iteratively generates and selects effective compressed prompts as task-specific demonstrations through style variation and in-context learning, enabling smaller models to act as efficient compressors with task-specific examples. Style-Compress outperforms two baseline compression models in four tasks: original prompt reconstruction, text summarization, multi-hop QA, and CoT reasoning. In addition, with only 10 samples and 100 queries for adaptation, prompts compressed by Style-Compress achieve performance on par with or better than original prompts at a compression ratio of 0.25 or 0.5.

Maram Hasanain, Fatema Ahmad, Firoj Alam

The use of propagandistic techniques in online content has increased in recent years aiming to manipulate online audiences. Fine-grained propaganda detection and extraction of textual spans where propaganda techniques are used, are essential for more informed content consumption. Automatic systems targeting the task over lower resourced languages are limited, usually obstructed by lack of large scale training datasets. Our study investigates whether Large Language Models (LLMs), such as GPT-4, can effectively extract propagandistic spans. We further study the potential of employing the model to collect more cost-effective annotations. Finally, we examine the effectiveness of labels provided by GPT-4 in training smaller language models for the task. The experiments are performed over a large-scale in-house manually annotated dataset. The results suggest that providing more annotation context to GPT-4 within prompts improves its performance compared to human annotators. Moreover, when serving as an expert annotator (consolidator), the model provides labels that have higher agreement with expert annotators, and lead to specialized models that achieve state-of-the-art over an unseen Arabic testing set. Finally, our work is the first to show the potential of utilizing LLMs to develop annotated datasets for propagandistic spans detection task prompting it with annotations from human annotators with limited expertise. All scripts and annotations will be shared with the community.

Jin Yea Jang, Saim Shin, Gahgene Gweon

The increasing use of AI agents in conversational services, such as counseling, highlights the importance of back-channeling (BC) as an active listening strategy to enhance conversational engagement. BC improves conversational engagement by providing timely acknowledgments and encouraging the speaker to talk. This study investigates the effect of BC provided by an AI agent on conversational engagement, offering insights for future AI conversational service design. We conducted an experiment with 55 participants, divided into Todak_BC and Todak_NoBC groups based on the presence or absence of the BC feature in Todak, a conversational agent. Each participant engaged in nine sessions with predetermined subjects and questions. We collected and analyzed approximately 6 hours and 30 minutes of conversation logs to evaluate conversational engagement using both quantitative (conversation persistence, including conversation duration and number of utterances) and qualitative metrics (context richness, including self-disclosure and topic diversity). The findings reveal significantly higher conversational engagement in the Todak_BC group compared to the Todak_NoBC group across all metrics (p<0.05). Additionally, the impact of BC varies across sessions, suggesting that conversation characteristics such as question type and topic sensitivity can influence BC effectiveness.

Wanqi Yang, Yanda Li, Meng Fang, Ling Chen

Time-Sensitive Question Answering (TSQA) demands the effective utilization of specific temporal contexts, encompassing multiple time-evolving facts, to address time-sensitive questions. This necessitates not only the parsing of temporal information within questions but also the identification and understanding of time-evolving facts to generate accurate answers. However, current large language models still have limited sensitivity to temporal information and their inadequate temporal reasoning capabilities. In this paper, we propose a novel framework that enhances temporal awareness and reasoning through Temporal Information-Aware Embedding and Granular Contrastive Reinforcement Learning. Experimental results on four TSQA datasets demonstrate that our framework significantly outperforms existing LLMs in TSQA tasks, marking a step forward in bridging the performance gap between machine and human temporal understanding and reasoning.

Pier Felice Balestrucci, Silvia Casola, Soda Marem Lo, Valerio Basile, Alessandro Mazzei

Generating ironic content is challenging: it requires a nuanced understanding of context and implicit references and balancing seriousness and playfulness. Moreover, irony is highly subjective and can depend on various factors, such as social, cultural, or generational aspects. This paper explores whether Large Language Models (LLMs) can learn to generate ironic responses to social media posts. To do so, we fine-tune two models to generate ironic and non-ironic content and deeply analyze their outputs’ linguistic characteristics, their connection to the original post, and their similarity to the human-written replies. We also conduct a large-scale human evaluation of the outputs. Additionally, we investigate whether LLMs can learn a form of irony tied to a generational perspective, with mixed results.

Nafis Sadeq, Zhouhang Xie, Byungkyu Kang, Prarit Lamba, Xiang Gao, Julian McAuley

Role-playing has wide-ranging applications in customer support, embodied agents, and computational social science. The influence of parametric world knowledge of large language models (LLMs) often causes role-playing characters to act out of character and to hallucinate about things outside the scope of their knowledge. In this work, we focus on the evaluation and mitigation of hallucination in fictional character role-play. We introduce a dataset with over 2,000 characters and 72,000 interviews, including 18,000 adversarial questions. We propose RoleFact, a role-playing method that mitigates hallucination by modulating the influence of parametric knowledge using a pre-calibrated confidence threshold. Experiments show that the proposed method improves the factual precision of generated responses by 18% for adversarial questions with a 44% reduction in temporal hallucination for time-sensitive interviews. The code and the dataset are available at https://github.com/NafisSadeq/rolefact.git.

Jian Li, Haojing Huang, Yujia Zhang, Pengfei Xu, Xi Chen, Rui Song, Lida Shi, Jingwen Wang, Hao Xu

Recently, there has been significant interest in replacing the reward model in Reinforcement Learning with Human Feedback (RLHF) methods for Large Language Models (LLMs), such as Direct Preference Optimization (DPO) and its variants. These approaches commonly use a binary cross-entropy mechanism on pairwise samples, i.e., minimizing and maximizing the loss based on preferred or dis-preferred responses, respectively. However, while this training strategy omits the reward model, it also overlooks the varying preference degrees within different responses. We hypothesize that this is a key factor hindering LLMs from sufficiently understanding human preferences. To address this problem, we propose a novel Self-supervised Preference Optimization (SPO) framework, which constructs a self-supervised preference degree loss combined with the alignment loss, thereby helping LLMs improve their ability to understand the degree of preference. Extensive experiments are conducted on two widely used datasets of different tasks. The results demonstrate that SPO can be seamlessly integrated with existing preference optimization methods and significantly boost their performance to achieve state-of-the-art performance. We also conduct detailed analyses to offer comprehensive insights into SPO, which verifies its effectiveness. The code is available at https://github.com/lijian16/SPO.

Ruirui Chen, Weifeng Jiang, Chengwei Qin, Ishaan Singh Rawal, Cheston Tan, Dongkyu Choi, Bo Xiong, Bo Ai

The important challenge of keeping knowledge in Large Language Models (LLMs) up-to-date has led to the development of various methods for incorporating new facts. However, existing methods for such knowledge editing still face difficulties with multi-hop questions that require accurate fact identification and sequential logical reasoning, particularly among numerous fact updates. To tackle these challenges, this paper introduces Graph Memory-based Editing for Large Language Models (GMeLLo), a straightforward and effective method that merges the explicit knowledge representation of Knowledge Graphs (KGs) with the linguistic flexibility of LLMs. Beyond merely leveraging LLMs for question answering, GMeLLo employs these models to convert free-form language into structured queries and fact triples, facilitating seamless interaction with KGs for rapid updates and precise multi-hop reasoning. Our results show that GMeLLo significantly surpasses current state-of-the-art (SOTA) knowledge editing methods in the multi-hop question answering benchmark, MQuAKE, especially in scenarios with extensive knowledge edits.

Erik Miehling, Manish Nagireddy, Prasanna Sattigeri, Elizabeth M. Daly, David Piorkowski, John T. Richards

Modern language models, while sophisticated, exhibit some inherent shortcomings, particularly in conversational settings. We claim that many of the observed shortcomings can be attributed to violation of one or more conversational principles. By drawing upon extensive research from both the social science and AI communities, we propose a set of maxims – quantity, quality, relevance, manner, benevolence, and transparency – for describing effective human-AI conversation. We first justify the applicability of the first four maxims (from Grice) in the context of human-AI interactions. We then argue that two new maxims, benevolence (concerning the generation of, and engagement with, harmful content) and transparency (concerning recognition of one’s knowledge boundaries, operational constraints, and intents), are necessary for addressing behavior unique to modern human-AI interactions. We evaluate the degree to which various language models are able to understand these maxims and find that models possess an internal prioritization of principles that can significantly impact accurate interpretability of the maxims.

Beiduo Chen, Xinpeng Wang, Siyao Peng, Robert Litschko, Anna Korhonen, Barbara Plank

Human label variation (HLV) is a valuable source of information that arises when multiple human annotators provide different labels for valid reasons. In Natural Language Inference (NLI) earlier approaches to capturing HLV involve either collecting annotations from many crowd workers to represent human judgment distribution (HJD) or use expert linguists to provide detailed explanations for their chosen labels. While the former method provides denser HJD information, obtaining it is resource-intensive. In contrast, the latter offers richer textual information but it is challenging to scale up to many human judges. Besides, large language models (LLMs) are increasingly used as evaluators (“LLM judges”) but with mixed results, and few works aim to study HJDs. This study proposes to exploit LLMs to approximate HJDs using a small number of expert labels and explanations. Our experiments show that a few explanations significantly improve LLMs’ ability to approximate HJDs with and without explicit labels, thereby providing a solution to scale up annotations for HJD. However, fine-tuning smaller soft-label aware models with the LLM-generated model judgment distributions (MJDs) presents partially inconsistent results: while similar in distance, their resulting fine-tuned models and visualized distributions differ substantially. We show the importance of complementing instance-level distance measures with a global-level shape metric and visualization to more effectively evaluate MJDs against human judgment distributions.

Ryuto Koike, Masahiro Kaneko, Naoaki Okazaki

To combat the misuse of Large Language Models (LLMs), many recent studies have presented LLM-generated-text detectors with promising performance. When users instruct LLMs to generate texts, the instruction can include different constraints depending on the user’s need. However, most recent studies do not cover such diverse instruction patterns when creating datasets for LLM detection. In this paper, we reveal that even task-oriented constraints — constraints that would naturally be included in an instruction and are not related to detection-evasion — cause existing powerful detectors to have a large variance in detection performance. We focus on student essay writing as a realistic domain and manually create task-oriented constraints based on several factors for essay quality. Our experiments show that the standard deviation (SD) of current detector performance on texts generated by an instruction with such a constraint is significantly larger (up to an SD of 14.4 F1-score) than that by generating texts multiple times or paraphrasing the instruction. We also observe an overall trend where the constraints can make LLM detection more challenging than without them. Finally, our analysis indicates that the high instruction-following ability of LLMs fosters the large impact of such constraints on detection performance.

Daria Galimzianova, Leonid Sanochkin

One of the main obstacles for deploying Active Learning (AL) in practical NLP tasks is high computational cost of modern deep learning models. This issue can be partially mitigated by applying lightweight models as an acquisition model, but it can lead to the acquisition-successor mismatch (ASM) problem. Previous works show that the ASM problem can be partially alleviated by using distilled versions of a successor models as acquisition ones. However, distilled versions of pretrained models are not always available. Also, the exact pipeline of model distillation that does not lead to the ASM problem is not clear. To address these issues, we propose to use adapters as an alternative to full fine-tuning for acquisition model training. Since adapters are lightweight, this approach reduces the training cost of the model. We provide empirical evidence that it does not cause the ASM problem and can help to deploy active learning in practical NLP tasks.

Daniil Moskovskiy, Sergey Pletenev, Alexander Panchenko

The lack of high-quality training data remains a significant challenge in NLP. Manual annotation methods, such as crowdsourcing, are costly, require intricate task design skills, and, if used incorrectly, may result in poor data quality. From the other hand, LLMs have demonstrated proficiency in many NLP tasks, including zero-shot and few-shot data annotation. However, they often struggle with text detoxification due to alignment constraints and fail to generate the required detoxified text. This work explores the potential of modern open source LLMs to annotate parallel data for text detoxification. Using the recent technique of activation patching, we generate a pseudo-parallel detoxification dataset based on ParaDetox. The detoxification model trained on our generated data shows comparable performance to the original dataset in automatic detoxification evaluation metrics and superior quality in manual evaluation and side-by-side comparisons.

Sara Vera Marjanovic, Haeun Yu, Pepa Atanasova, Maria Maistro, Christina Lioma, Isabelle Augenstein

Knowledge-intensive language understanding tasks require Language Models (LMs) to integrate relevant context, mitigating their inherent weaknesses, such as incomplete or outdated knowledge. However, conflicting knowledge can be present in the LM’s parameters, termed intra-memory conflict, which can affect a model’s propensity to accept contextual knowledge. To study the effect of intra-memory conflict on LM’s ability to accept the relevant context, we utilise two knowledge conflict measures and a novel dataset containing inherently conflicting data, DYNAMICQA. This dataset includes facts with a temporal dynamic nature where facts can change over time and disputable dynamic facts, which can change depending on the viewpoint. DYNAMICQA is the first to include real-world knowledge conflicts and provide context to study the link between the different types of knowledge conflicts. We also evaluate several measures on their ability to reflect the presence of intra-memory conflict: semantic entropy and a novel coherent persuasion score. With our extensive experiments, we verify that LMs show a greater degree of intra-memory conflict with dynamic facts compared to facts that have a single truth value. Further, we reveal that facts with intra-memory conflict are harder to update with context, suggesting that retrieval-augmented generation will struggle with the most commonly adapted facts