论文检索

Wei Sun, Tingyu Qu, Mingxiao Li, Jesse Davis, Marie-Francine Moens

Efficiently updating multilingual knowledge in large language models (LLMs) without disrupting coherent factual representations across languages remains a significant challenge. While deploying separate editing systems for each language might seem viable, this approach incurs substantial costs due to the need to manage multiple models. A more efficient solution involves integrating knowledge updates across all languages into a unified model. However, sequential edits across languages often lead to destructive parameter interference, significantly degrading multilingual generalization and the accuracy of injected knowledge. To address this issue, we propose LangEdit, a novel null-space constrained framework designed to precisely isolate language-specific knowledge updates. The core innovation of LangEdit lies in its ability to project parameter updates for each language onto the orthogonal complement of other languages’ subspaces. This approach mathematically guarantees update independence while preserving multilingual generalization capabilities. We conduct a comprehensive evaluation across three model architectures, six languages, and four downstream tasks, demonstrating that LangEdit effectively mitigates parameter interference and outperforms existing state-of-the-art editing methods. Our results highlight its potential for enabling efficient and accurate multilingual knowledge updates in LLMs.

Yuan Sui, Yufei He, Nian Liu, Xiaoxin He, Kun Wang, Bryan Hooi

Large Language Models (LLMs) are often challenged by generating erroneous or hallucinated responses, especially in complex reasoning tasks. Leveraging Knowledge Graphs (KGs) as external knowledge sources has emerged as a viable solution. However, existing KG-enhanced methods, either retrieval-based or agent-based, encounter difficulties in accurately retrieving knowledge and efficiently traversing KGs at scale. In this paper, we propose a unified framework, FiDeLiS, designed to improve the factuality of LLM responses by anchoring answers to verifiable reasoning steps retrieved from KGs. To achieve this, we leverage step-wise beam search with a deductive scoring function, allowing the LLM to validate reasoning process step by step, and halt the search once the question is deducible. In addition, we propose a Path-RAG module to pre-select a smaller candidate set for each beam search step, reducing computational costs by narrowing the search space. Extensive experiments show that our method, as a training-free framework, not only improve the performance but also enhance the factuality and interpretability across different benchmarks.

Qitao Qin, Yucong Luo, Yihang Lu, Zhibo Chu, Xiaoman Liu, Xianwei Meng

Retrieval-Augmented Generation (RAG), by integrating non-parametric knowledge from external knowledge bases into models, has emerged as a promising approach to enhancing response accuracy while mitigating factual errors and hallucinations. This method has been widely applied in tasks such as Question Answering (QA). However, existing RAG methods struggle with open-domain QA tasks because they perform independent retrieval operations and directly incorporate the retrieved information into generation without maintaining a summarizing memory or using adaptive retrieval strategies, leading to noise from redundant information and insufficient information integration.To address these challenges, we propose Adaptive memory-based optimization for enhanced RAG (Amber) for open-domain QA tasks, which comprises an Agent-based Memory Updater, an Adaptive Information Collector, and a Multi-granular Content Filter, working together within an iterative memory updating paradigm. Specifically, Amber integrates and optimizes the language model’s memory through a multi-agent collaborative approach, ensuring comprehensive knowledge integration from previous retrieval steps. It dynamically adjusts retrieval queries and decides when to stop retrieval based on the accumulated knowledge, enhancing retrieval efficiency and effectiveness. Additionally, it reduces noise by filtering irrelevant content at multiple levels, retaining essential information to improve overall model performance. We conduct extensive experiments on several open-domain QA datasets, and the results demonstrate the superiority and effectiveness of our method and its components. The source code is available .

Xiaoxue Cheng, Junyi Li, Wayne Xin Zhao, Ji-Rong Wen

Large language models (LLMs) demonstrate exceptional capabilities, yet still face the hallucination issue. Typical text generation approaches adopt an auto-regressive generation without deliberate reasoning, often leading to untrustworthy and factually inaccurate responses. In this paper, we propose HaluSearch, a novel framework that incorporates tree search-based algorithms (e.g., MCTS) to enable an explicit slow thinking generation process for mitigating hallucinations during inference. Specifically, HaluSearch frames text generation as a step-by-step reasoning process, using a self-evaluation reward model to score each generation step and guide the tree search towards the most reliable generation pathway. To balance efficiency and quality, we introduce a hierarchical system switch mechanism, which dynamically switches between fast and slow thinking modes at both instance and step levels. We conduct extensive experiments on both English and Chinese datasets, and the results show that our approach significantly outperforms baseline approaches.

Xinyan Jiang, Hang Ye, Yongxin Zhu, Xiaoying Zheng, Zikang Chen, Jun Gong

Large Language Models (LLMs) often generate hallucinations, producing outputs that are contextually inaccurate or factually incorrect. We introduce HICD, a novel method designed to induce hallucinations for contrastive decoding to mitigate hallucinations. Unlike existing contrastive decoding methods, HICD selects attention heads crucial to the model’s prediction as inducing heads, then induces hallucinations by dispersing attention of these inducing heads and compares the hallucinated outputs with the original outputs to obtain the final result. Our approach significantly improves performance on tasks requiring contextual faithfulness, such as context completion, reading comprehension, and question answering. It also improves factuality in tasks requiring accurate knowledge recall. We demonstrate that our inducing heads selection and attention dispersion method leads to more “contrast-effective” hallucinations for contrastive decoding, outperforming other hallucination-inducing methods. Our findings provide a promising strategy for reducing hallucinations by inducing hallucinations in a controlled manner, enhancing the performance of LLMs in a wide range of tasks.

Bowen Ping, Jiali Zeng, Fandong Meng, Shuo Wang, Jie Zhou, Shanghang Zhang

Recent advancements in large language models (LLMs) have markedly improved their capacity to handle long text inputs; however, current models, including GPT-4o, still exhibit unsatisfactory performance in long-form generation. Generating high-quality long-form content still remains a significant challenge. In this paper, we present LongDPO, a novel approach designed to enhance long-form text generation through step-level supervision. By leveraging Monte Carlo Tree Search (MCTS) to collect stepwise preference pairs and employing a global memory pool to maintain factual accuracy, LongDPO effectively mitigates issues such as inconsistencies that are prevalent in long-context LLMs. Furthermore, we integrate critique-augmented generation to refine the selected preference pairs. Following the collection of stepwise preference pairs, we apply stepwise preference learning for fine-grained optimization. Experimental results demonstrate that our method enhances performance on long-form generation benchmarks (e.g. LongBench-Write) while maintaining nearly lossless performance on several general benchmarks.

Sirui Huang, Yanggan Gu, Zhonghao Li, Xuming Hu, Li Qing, Guandong Xu

Large language models (LLMs) have made significant strides in natural language processing by leveraging their ability to comprehend and reason with factual knowledge. However, a significant amount of factual knowledge is stored in structured data, which has unique characteristics not typically encountered in the unstructured texts used for pretraining LLMs. To evaluate the capability of LLMs in handling facts structurally stored, we introduce a benchmark called StructFact, which includes meticulously annotated factual questions, spanning five tasks that reflect the intrinsic properties of structured data. This benchmark aims to delineate the strengths and limitations of LLMs in reasoning with structured data for knowledge-intensive tasks in practical applications. Extensive experiments conducted on 10 common LLMs have yielded several insights, one notable finding being that these models struggle significantly with the heterogeneity of structured data during reasoning.

Litu Ou, Mirella Lapata

Hierarchical Merging is a technique commonly used to summarize very long texts (>100K tokens) by breaking down the input into smaller sections, summarizing those sections individually, and then merging or combining those summaries into a final coherent summary. Although it helps address the limitations of large language models (LLMs) with fixed input length constraints, the recursive merging process can amplify LLM hallucinations, increasing the risk of factual inaccuracies. In this paper, we seek to mitigate hallucinations by enriching hierarchical merging with context from the source document. Specifically, we propose different approaches to contextual augmentation ranging from *replacing* intermediate summaries with relevant input context, to *refining* them while using the context as supporting evidence, and *aligning* them implicitly (via citations) to the input. Experimental results on datasets representing legal and narrative domains show that contextual augmentation consistently outperforms zero-shot and hierarchical merging baselines for the Llama 3.1 model family. Our analysis further reveals that refinement methods tend to perform best when paired with extractive summarization for identifying relevant input.

Qing Zong, Zhaowei Wang, Tianshi Zheng, Xiyu Ren, Yangqiu Song

The rapid development of LLMs has sparked extensive research into their factual knowledge. Current works find that LLMs fall short on questions around low-frequency entities. However, such proofs are unreliable since the questions can differ not only in entity frequency but also in difficulty themselves. So we introduce **ComparisonQA** benchmark, containing **283K** abstract questions, each instantiated by a pair of high-frequency and low-frequency entities. It ensures a controllable comparison to study the role of knowledge frequency in the performance of LLMs. Because the difference between such a pair is only the entity with different frequencies. In addition, we use both correctness and uncertainty to develop a two-round method to evaluate LLMs’ knowledge robustness. It aims to avoid possible semantic shortcuts which is a serious problem of current QA study. Experiments reveal that LLMs, including GPT-4o, exhibit particularly low robustness regarding low-frequency knowledge. Besides, we find that uncertainty can be used to effectively identify high-quality and shortcut-free questions while maintaining the data size. Based on this, we propose an automatic method to select such questions to form a subset called **ComparisonQA-Hard**, containing only hard low-frequency questions.

Zihang Liu, Jiawei Guo, Hao Zhang, Hongyang Chen, Jiajun Bu, Haishuai Wang

While Large Language Models (LLMs) have exhibited impressive performance in generating long-form content, they frequently present a hazard of producing factual inaccuracies or hallucinations. An effective strategy to mitigate this hazard is to leverage off-the-shelf LLMs to detect hallucinations after the generation. The primary challenge resides in the comprehensive elicitation of the intrinsic knowledge acquired during their pre-training phase. However, existing methods that employ multi-step reasoning chains predominantly fall short of addressing this issue. Moreover, since existing methods for hallucination detection tend to decompose text into isolated statements, they are unable to understand the contextual semantic relations in long-form content. In this paper, we study a novel concept, self-elicitation, to leverage self-generated thoughts derived from prior statements as catalysts to elicit the expression of intrinsic knowledge and understand contextual semantics. We present a framework, SelfElicit, to integrate self-elicitation with graph structures to effectively organize the elicited knowledge and facilitate factual evaluations. Extensive experiments on five datasets in various domains demonstrate the effectiveness of self-elicitation and the superiority of our proposed method.

Weijie Shi, Hao Chen, Jiaming Li, Yao Zhao, Yazhong Zhang, Qijin Chen, Jipeng Zhang, Ruiyuan Zhang, Jia Zhu, Jiajie Xu 等

Retrieval-augmented language models (RALMs) aim to incorporate external knowledge to address the issues of factual hallucination and knowledge obsolescence faced by large language models (LLMs). Inevitably, the retrieved passages based on similarity search may be irrelevant to the given question, and the aggregation of these passages can confuse the model to give a correct answer. To improve the performance of RALM in such conditions, we propose layer-knowledge guided attention for RALMs, which harnesses the layer-wise knowledge of LLMs to optimize per-layer attention on useful passages, making the model pay attention to the most relevant content and ignore irrelevant ones. Specifically, we first systematically study LLM’s attention patterns and their relationship with the accuracy of RALM responses, where middle-focus attentions play a crucial role in selectively gathering relevant information. Based on this, a layer-wise passage estimator leverages the varied knowledge encoded across LLM layers to assess not only passage relevance scores but also associated confidences. Finally, a relevance-aware passage fusion enables selective attention to relevant passages, mitigating distractibility and positional bias of causal attention. Experiments show that our method outperforms existing methods on RALM benchmarks.

James Xu Zhao, Jimmy Z.J. Liu, Bryan Hooi, See-Kiong Ng

Large language models (LLMs) are widely used for long-form text generation. However, factual errors in the responses would undermine their reliability. Despite growing attention to LLM factuality, the effect of response length on factuality remains underexplored. In this work, we systematically investigate this relationship by first introducing an automatic and bi-level long-form factuality evaluation framework, which achieves high agreement with human annotations while being cost-effective. Using this framework, we conduct controlled experiments and find that longer responses exhibit lower factual precision, confirming the presence of length bias. To explain this phenomenon, we empirically examine three hypotheses: error propagation, long context, and facts exhaustion. Our results reveal that facts exhaustion, where the model gradually exhausts more reliable knowledge, is the primary cause of factual degradation, rather than the other two hypotheses.

Soo Kyung Kim, Hyunsoo Cho

Large Language Models (LLMs) often succumb to adversarial prompts, a phenomenon popularly known as “jailbreaking.” While jailbreaking primarily targets short-term noncompliance with predefined policies, we argue that a deeper vulnerability lies in altering an LLM’s fundamental axiomatic beliefs, such as mathematical or philosophical truths. In this work, we introduce GoodLiar, a reinforcement learning (RL)-based framework that generates deceptive contexts to systematically rewrite an LLM’s core logical or philosophical understandings. By incentivizing an RL agent to produce persuasive and coherent arguments, GoodLiar aims to induce persistent belief shifts, rather than merely influencing immediate judgments of factual truthfulness. %rather than one-off policy breaches. Our approach introduces DA-ILQL, a novel offline RL method that extends ILQL by integrating on-policy data and language exploration to enhance the language discovery and optimization. Through extensive evaluations on multiple LLMs, we show that deceptive contexts discovered by GoodLiar consistently outperform simple multi-turn prompting methods.

Chuanyuan Tan, Wenbiao Shao, Hao Xiong, Tong Zhu, Zhenhua Liu, Kai Shi, Wenliang Chen

Handling unanswerable questions (UAQ) is crucial for LLMs, as it helps prevent misleading responses in complex situations. While previous studies have built several datasets to assess LLMs’ performance on UAQ, these datasets lack factual knowledge support, which limits the evaluation of LLMs’ ability to utilize their factual knowledge when handling UAQ. To address the limitation, we introduce a new unanswerable question dataset UAQFact, a bilingual dataset with auxiliary factual knowledge created from a Knowledge Graph. Based on UAQFact, we further define two new tasks to measure LLMs’ ability to utilize internal and external factual knowledge, respectively. Our experimental results across multiple LLM series show that UAQFact presents significant challenges, as LLMs do not consistently perform well even when they have factual knowledge stored. Additionally, we find that incorporating external knowledge may enhance performance, but LLMs still cannot make full use of the knowledge which may result in incorrect responses. Our code and dataset are available at https://github.com/cytan17726/UAQ_Fact.

Korbinian Q. Weidinger, Santosh T.Y.S.S, Oana Ichim, Matthias Grabmair

LLMs have become prevalent tools for information seeking across various fields, including law. However, their generated responses often suffer from hallucinations, hindering their widespread adoption in high stakes domains such as law, which can potentially mislead experts and propagate societal harms. To enhance trustworthiness in these systems, one promising approach is to attribute the answer to an actual source, thereby improving the factuality and verifiability of the response. In pursuit of advancing attributed legal question answering, we introduce AQuAECHR, a benchmark comprising information-seeking questions from ECHR jurisprudence along with attributions to relevant judgments. We present strategies to automatically curate this dataset from ECHR case law guides and utilize an LLM-based filtering pipeline to improve dataset quality, as validated by legal experts. Additionally, we assess several LLMs, including those trained on legal corpora, on this dataset to underscore significant challenges with the current models and strategies dealing with attributed QA, both quantitatively and qualitatively.

Guocong Li, Weize Liu, Yihang Wu, Ping Wang, Shuaihan Huang, Hongxia Xu, Jian Wu

Large language models (LLMs) exhibit excellent performance in natural language processing (NLP), but remain highly sensitive to the quality of input queries, especially when these queries contain misleading or inaccurate information. Existing methods focus on correcting the output, but they often overlook the potential of improving the ability of LLMs to detect and correct misleading content in the input itself. In this paper, we propose a novel three-stage fine-tuning method that enhances the ability of LLMs to detect and correct misleading information in the input, further improving response accuracy and reducing hallucinations. Specifically, the three stages include (1) training LLMs to identify misleading information, (2) training LLMs to correct the misleading information using built-in or external knowledge, and (3) training LLMs to generate accurate answers based on the corrected queries. To evaluate our method, we conducted experiments on three datasets for the hallucination detection task and the question answering (QA) task, as well as two datasets containing misleading information that we constructed. The experimental results demonstrate that our method significantly improves the accuracy and factuality of LLM responses, while also enhancing the ability to detect hallucinations and reducing the generation of hallucinations in the output, particularly when the query contains misleading information.

Kun Zhang, Oana Balalau, Ioana Manolescu

Analysing the differences in how events are represented across texts, or verifying whether the language model generations hallucinate, requires the ability to systematically compare their content. To support such comparison, structured representation that captures fine-grained information plays a vital role.In particular, identifying distinct atomic facts and the discourse relations connecting them enables deeper semantic comparison. Our proposed approach combines structured discourse information extraction with a classifier, FDSpotter, for factual consistency verification. We show that adversarial discourse relations pose challenges for language models, but fine-tuning on our annotated data, DiscInfer, achieves competitive performance. Our proposed approach advances factual consistency verification by grounding in linguistic structure and decomposing it into interpretable components. We demonstrate the effectiveness of our method on the evaluation of two tasks: data-to-text generation and text summarisation. Our code and dataset will be publicly available on GitHub.

Zain Muhammad Mujahid, Dilshod Azizov, Maha Tufail Agro, Preslav Nakov

In an age characterized by the proliferation of mis- and disinformation online, it is critical to empower readers to understand the content they are reading. Important efforts in this direction rely on manual or automatic fact-checking, which can be challenging for emerging claims with limited information. Such scenarios can be handled by assessing the reliability and the political bias of the source of the claim, i.e., characterizing entire news outlets rather than individual claims or articles. This is an important but understudied research direction. While prior work has looked into linguistic and social contexts, we do not analyze individual articles or information in social media. Instead, we propose a novel methodology that emulates the criteria that professional fact-checkers use to assess the factuality and political bias of an entire outlet. Specifically, we design a variety of prompts based on these criteria and elicit responses from large language models (LLMs), which we aggregate to make predictions. In addition to demonstrating sizable improvements over strong baselines via extensive experiments with multiple LLMs, we provide an in-depth error analysis of the effect of media popularity and region on model performance. Further, we conduct an ablation study to highlight the key components of our dataset that contribute to these improvements. To facilitate future research, we released our dataset and code.

Lingyuan Liu, Mengxiang Zhang

Large language models (LLMs)-based query expansion for information retrieval augments queries with generated hypothetical documents with LLMs. However, its performance relies heavily on the scale of the language models (LMs), necessitating larger, more advanced LLMs. This approach is costly, computationally intensive, and often has limited accessibility. To address these limitations, we introduce GOLFer - Smaller LMs-Generated Documents Hallucination Filter & Combiner - a novel method leveraging smaller open-source LMs for query expansion. GOLFer comprises two modules: a hallucination filter and a documents combiner. The former detects and removes non-factual and inconsistent sentences in generated documents, a common issue with smaller LMs, while the latter combines the filtered content with the query using a weight vector to balance their influence. We evaluate GOLFer alongside dominant LLMs-based query expansion methods on three web search and ten low-resource datasets. Experimental results demonstrate that GOLFer consistently outperforms other methods using smaller LMs, and maintains competitive performance against methods using large-size LLMs, demonstrating its effectiveness.

Chengwei Qin, Wenxuan Zhou, Karthik Abinav Sankararaman, Nanshu Wang, Tengyu Xu, Alexander Radovic, Eryk Helenowski, Arya Talebzadeh, Aditya Tayade, Sinong Wang 等

Hallucination, the generation of factually incorrect information, remains a significant challenge for large language models (LLMs), especially in open-domain long-form generation. Existing approaches for detecting hallucination in long-form tasks either focus on limited domains or rely heavily on external fact-checking tools, which may not always be available.In this work, we systematically investigate reference-free hallucination detection in open-domain long-form responses. Our findings reveal that internal states (e.g., model’s output probability and entropy) alone are insufficient for reliably (i.e., better than random guessing) distinguishing between factual and hallucinated content. To enhance detection, we explore various existing approaches, including prompting-based methods, probing, and fine-tuning, with fine-tuning proving the most effective. To further improve the accuracy, we introduce a new paradigm, named RATE-FT, that augments fine-tuning with an auxiliary task for the model to jointly learn with the main task of hallucination detection. With extensive experiments and analysis using a variety of model families & datasets, we demonstrate the effectiveness and generalizability of our method, e.g., +3% over general fine-tuning methods on LongFact.