论文检索

Fuda Ye, Jiachuan Wang, Yongqi Zhang, Lei Chen, Shuangyin Li

Retrieval-augmented generation (RAG) extends the capabilities of large language models (LLMs) by providing access to external knowledge. However, traditional retrieval-augmented LLMs rely on a silent reading paradigm that processes all retrieved documents passively, forcing them to reason without any interaction with the documents. This paradigm contrasts sharply with human interactive reading behavior, where external tools, such as bookmarks and notes, are used to offload cognitive demands. This paper introduces BubbleRAG, an enhanced RAG framework that emulates human interactive reading through annotation and re-reading. Specifically, BubbleRAG utilizes a lightweight thought bubble module that offloads LLM’s internal cognition into external bookmark tokens, which are then annotated back into the context. These bookmarks serve as externalized memory, allowing the LLM to revisit these annotations in subsequent reading and answering. Notably, BubbleRAG is particularly suitable for low-resource scenarios, as the LLM parameters remain frozen. Extensive experiments confirm the effectiveness, robustness, and generalizability of BubbleRAG. Our findings demonstrate that BubbleRAG enables LLMs to achieve superior evidence identification abilities typically seen in retrievers, while establishing a cognitive link between external and internal information during answer generation. The source code is available at https://github.com/yefd/BubbleRAG.

Jinlong Ma, Yu Zhang, Xuefeng Bai, Kehai Chen, Yuwei Wang, Zeming Liu, Jun Yu, Min Zhang

Grounded Multimodal Named Entity Recognition (GMNER) aims to extract text-based entities, assign them semantic categories, and ground them to corresponding visual regions. In this work, we explore the potential of Multimodal Large Language Models (MLLMs) to perform GMNER in an end-to-end manner, moving beyond their typical role as auxiliary tools within cascaded pipelines.Crucially, our investigation reveals a fundamental challenge: MLLMs exhibit \textit{modality bias}, including visual bias and textual bias, which stems from their tendency to take unimodal shortcuts rather than rigorous cross-modal verification.To address this, we propose Modality-aware Consistency Reasoning (\textbf{MCR}), which enforces structured cross-modal reasoning through Multi-style Reasoning Schema Injection (MRSI) and Constraint-guided Verifiable Optimization (CVO). MRSI transforms abstract constraints into executable reasoning chains, while CVO empowers the model to dynamically align its reasoning trajectories with Group Relative Policy Optimization (GRPO).Experiments on GMNER and visual grounding tasks demonstrate that MCR effectively mitigates modality bias and achieves superior performance compared to existing baselines.

Ofek Raban, Gal Chechik, Ethan Fetaya

该来源暂未提供摘要。

Nicholas Popovič, Michael Färber

We study how large language models recall relational knowledge during text generation, with a focus on identifying latent representations suitable for relation classification via linear probes.Prior work shows how attention heads and MLPs interact to resolve subject, predicate, and object, but it remains unclear which representations support faithful linear relation classification and why some relation types are easier to capture linearly than others.We systematically evaluate different latent representations derived from attention head and MLP contributions, showing that per-head attention contributions to the residual stream are comparatively strong features for linear relation classification.Feature attribution analyses of the trained probes, as well as characteristics of the different relation types, reveal clear correlations between probe accuracy and relation specificity, entity connectedness, and how distributed the signal on which the probe relies is across attention heads.Finally, we show how token-level feature attribution of probe predictions can be used to reveal probe behavior in further detail.

Girish, Mohd Mujtaba Akhtar, Orchid Chetia Phukan, Arun Balaji Buduru

The rapid advancement of Audio Large Language Models (ALMs), driven by Neural Audio Codecs (NACs), has led to the emergence of highly realistic speech deepfakes, commonly referred to as CodecFakes (CFs). Consequently, CF detection has attracted increasing attention from the research community. However, existing studies predominantly focus on English or Chinese, leaving the vulnerability of Indic languages largely unexplored. To bridge this gap, we introduce Indic-CodecFake (ICF) dataset, the first large-scale benchmark comprising real and NAC-synthesized speech across multiple Indic languages, diverse speaker profiles, and multiple NAC types. We use IndicSUPERB as the real speech corpus for generation of ICF dataset. Our experiments demonstrate that state-of-the-art (SOTA) CF detectors trained on English-centric datasets fail to generalize to ICF, underscoring the challenges posed by phonetic diversity and prosodic variability in Indic speech. Further, we present systematic evaluation of SOTA ALMs in a zero-shot setting on ICF dataset. We evaluate these ALMs as they have shown effectiveness for different speech tasks. However, our findings reveal that current ALMs exhibit consistently poor performance. To address this, we propose SATYAM, a novel hyperbolic ALM tailored for CF detection in Indic languages. SATYAM integrates semantic representations from Whisper and prosodic representations from TRILLsson using through Bhattacharya distance in hyperbolic space, and subsequently performs the same alignment procedure between the fused speech representation and a input conditioning prompt. This dual-stage fusion framework enables SATYAM to effectively model hierarchical relationships both within speech (semantic–prosodic) and across modalities (speech–text). Extensive evaluations show that SATYAM consistently outperforms competitive end-to-end and ALM-based baselines on the ICF benchmark.

Khashayar Alavi, Zhastay Yeltay, Lucie Flek, Akbar Karimi

When LLM agents work together, they seem to be more powerful than a single LLM in mathematical question answering. However, are they also more robust to adversarial inputs? We investigate this question using adversarially perturbed math questions. These perturbations include punctuation noise with three intensities (10%, 30%, 50%), plus real-world and human-like typos (WikiTypo, R2ATA). Using a unified sampling-and-voting framework (Agent Forest), we evaluate six open-source models (Qwen3-4B/14B, Llama3.1-8B, Mistral-7B, Gemma3-4B/12B) across four benchmarks (GSM8K, MATH, MMLU–Math, MultiArith), with various numbers of agents n = 1,2,5,10,15,20,25. Our findings show that 1) Noise type matters: punctuation noise harm scales with its severity, and the human typos remain the dominant bottleneck, yielding the largest gaps to Clean accuracy and the highest attack success rate (ASR) even with a large number of agents; 2) Collaboration reliably improves accuracy as the number of agents, n, increases, with the largest gains from n=1 to n=5 and diminishing returns beyond n\approx10. However, the adversarial robustness gap persists regardless of the agent count.

Lichang Song, Ting Long, Yi Chang

Retrieval-Augmented Generation (RAG) has demonstrated strong effectiveness in knowledge-intensive tasks by grounding language generation in external evidence. Despite its success, many existing RAG systems are built based on a ranking-centric, asymmetric dependency paradigm, where the generation quality of the generator is highly dependent on reranking results of the reranker.To overcome this limitation, we propose Cooperative Retrieval-Augmented Generation (CoRAG), a framework that treats the reranker and the generator as peer decision-makers rather than being connected through an asymmetric dependency pipeline. By jointly optimizing their behaviors toward a shared task objective, the reranker and generator are encouraged to cooperate, ensuring that document reranking and generation work in concert to improve the final response.Experimental results demonstrate good generalization and improved generation stability of CoRAG, even when the model is trained on only around 10K PopQA samples. Our model released in https://github.com/CoderrrSong/CoRAG

Syed Ahad, Burhanuddin Aliasghar Ezzi, Muhammad Arsalan Hussain, Sandesh Kumar, Abdul Samad

Grammatical Error Correction (GEC) for Urdu remains an under-researched area due to the lack of annotated datasets. This paper addresses the challenge of generating a robust corpus for fine-tuning deep learning models aimed at Urdu GEC. We propose a method for synthesizing a large dataset by collecting errors from the Urdu WikiEdits history, learning from them, and inserting similar errors in grammatically correct sentences to generate incorrect sentences with grammatical errors, hence creating a pair of grammatically correct and incorrect sentences. We introduce UrduGEC-Synthetic, a synthetically generated dataset produced through this pipeline. Furthermore, we introduce UrduGEC-Gold, a Gold Dataset by extracting errors from exam copies of students. Finally, we also fine-tuned various models on UrduGEC-Synthetic and evaluated them against UrduGEC-Gold to show the quality of synthetic data generation.

Yifan Wu, Jingze Shi, Bingheng Wu, Jiayi Zhang, Xiaotian Lin, Yizhang Zhu, Zhaoyang Yu, Bang Liu, Chenglin Wu, Nan Tang 等

Human experts tackle difficult math problems by identifying and executing a few pivotal steps rather than listing every intermediate thought. In contrast, standard Chain-of-Thought (CoT) distillation trains small models on lengthy reasoning traces, encouraging a uniform overthinking style across easy and hard items alike. The result is rigid, slow solutions that sacrifice adaptivity. This approach stands in sharp contrast to human intuition. Humans naturally adapt their problem-solving strategy, dedicating significant effort to difficult problems while finding quick, simple solutions for easier ones. We argue that the root cause lies in the training data: it contains excess information and reasoning steps organized in ways misaligned with human practice. We address this with Difficulty-Aware Distillation(DAD), a procedure for producing training data that mirrors concise human reasoning. A large teacher model first assesses a problem’s difficulty and then rewrites the solution to retain only the essential steps. Using this process, we constructed LiteCoT, a 100,000-example corpus of short, clear rationales, and used it to train our Liter models. With 100k LiteCoT, we outperform models trained on 800k long CoT and cut both training and inference costs. The advantage is consistent across standard math benchmarks, showing that concise, human-aligned data delivers equal or better accuracy with much less compute. For example, on the challenging AIME24 exam, our approach reaches 74.2% Pass@1 using only about 5K inference tokens, surpassing other methods that consume many more tokens.

Wenqiang Wang, Peng Chen, Yan Xiao, Yangshijie Zhang, Xiaoyue Lu, Jianjie Huang, Xiaochun Cao

Standard in-context learning (ICL) assumes identical output spaces between test and retrieval datasets (fully aligned). However, in practice, these datasets can be fully aligned, partially aligned, or fully disjoint in label space (Output space), forming an information continuum from rich to scarce. Naive ICL often becomes ineffective under such mismatches. In this work, we challenge this assumption by demonstrating that the retrieval dataset need not perfectly align with the test dataset, as long as it remains related to the target task. We propose Task-Related In-Context Learning (TRICL), a unified framework for ICL under output-space mismatch, designed to cover the full continuum of scenarios. TRICL first identifies demonstrations in the mismatched retrieval dataset that are relevant to the test label space via a lightweight Bayesian probabilistic criterion, and uses them to form a related dataset. TRICL then perform ICL on the related dataset to obtain preliminary predictions; finally, TRICL leverage these intermediate predictions to reduce and transform the output space of the original test task, thereby improving the performance of LLMs. Even in the most information-scarce fully disjoint scenario, as long as the retrieval dataset is task-related to the test task, TRICL achieves state-of-the-art (SOTA) results across three LLMs, three task types, and four datasets. Moreover, TRICL remains effective in the fully aligned and partially aligned scenarios, consistently yielding strong gains over competitive baselines. Moreover, TRICL also extends to generative task.

Jungyoub Cha, Hyunjong Kim, Sungzoon Cho

Speculative decoding is a widely used technique for accelerating inference in large language models (LLMs), but its performance degrades as input length grows, with significant drops even at moderate lengths. Yet, this early degradation has remained largely underexplored. We introduce SpecExtend, a drop-in enhancement that improves speculative decoding on long sequences without additional training. SpecExtend integrates efficient attention mechanisms such as FlashAttention and Hybrid Tree Attention to accelerate prefill and verification steps. To improve both draft accuracy and speed on long inputs without retraining, we propose Cross-model Retrieval, a novel KV cache eviction strategy that leverages the target model’s attention scores to dynamically select relevant context for the smaller draft model. Extensive evaluations show that SpecExtend accelerates speculative decoding by up to 2.84× on 16K-token long document summarization and up to 3.86× on long-form reasoning, while preserving the short-input performance of state-of-the-art frameworks.

Mingxu Tao, Jiawei Hu, Xian Zhou, Wenpeng Hu, Jiajun Cheng, Yunbo Cao, Zhunchen Luo, Guotong Geng

Legal case retrieval remains challenging due to the complexity of legal language and the need for precise lexical alignment between queries and relevant cases. Although dense retrieval models have achieved notable progress, empirical studies show that BM25 continues to serve as a strong baseline in this domain. It motivates us to propose a self-evolving framework for rule-driven query rewriting that enhances BM25 without any parameter training. The framework equips an LLM-based agent with an automatic evaluation environment, enabling it to iteratively create rewriting rules, plan validation experiments over rule combinations, and eliminate ineffective rules based on historical feedbacks. We evaluate our method on the Chinese legal case retrieval benchmark LeCaRD-v2. Experimental results demonstrate that the proposed framework outperforms non-evolutionary baselines, including human-designed rules and greedy rule selection, particularly when powered by a high-capacity core LLM. We also conduct detailed analyses to investigate the mechanisms underlying self-evolution. Our findings reveal that LLM’s capabilities to leverage previous experimental results and its intrinsic knowledge of rule elimination play critical roles in refining the rule set via self-evolution.

Yu Zou, Yan Chen, Lida He, Qi Zhou, Xiaorui Zhou, Aixi Zhong, Yi Wang, Wei Li, Qingyu Wang, Jiatao Li 等

Understanding financial documents is critical for high-stakes decision-making yet hindered by systemic semantic implicitness: key facts are rarely explicit in surface text and often determined by global structural cues. Missing these cues invites semantic misinterpretations, such as misreading what a number refers to, an outcome unacceptable in high-stakes environments. However, existing Retrieval-Augmented Generation (RAG) systems typically treat structure as a physical navigational skeleton rather than intrinsic semantic knowledge. To address this, we introduce Fin-STAR (Financial STructure-As-Semantics Retrieval), a framework redefining hierarchy as intrinsic semantics. Fin-STAR incorporates a novel Structure-Enriched Semantic Indexing mechanism that augments the hierarchical lineage with snippet-derived virtual nodes, and injects this enriched context via a semantic cross-attention paradigm, rendering implicit cues explicit. By grounding evidence within its structural scope, we preserve factual invariance and ensure contextual integrity. Addressing the lack of granular public datasets, we conduct experiments on FinTierQA Gold, a curated expert benchmark. Results show that Fin-STAR outperforms state-of-the-art hierarchical and graph-based baselines across diverse query complexities, document types, and markets. Notably, ablations confirm that our semantic injection consistently outperforms alternative strategies. Finally, we release FinTierQA, comprising 3.9M pairs automatically constructed from 78k documents via our framework .

Xiao Xia, Dan Zhang, Tianrui Sun

Providing accurate reward signals for code generated by large language models (LLMs) is a significant challenge in applying reinforcement learning (RL) to code generation. Existing methods rely on unit tests to evaluate code correctness and provide rewards, which are hindered by the difficulty of acquiring and verifying reliable unit tests at scale. In this work, we propose CodeRM-NT, a code reward model with no reliance on unit tests. Our method leverages Monte Carlo Tree Search guided by LLMs to generate code snippets and judges execution traces to annotate code with reward signals. We use the rewards to train CodeRM-NT that is capable of providing rewards for code during RL. CodeRM-NT also facilitates curriculum learning by scoring and sorting training samples based on their difficulty. Experimental results demonstrate that training with CodeRM-NT consistently outperforms synthetic unit test-based rewards, yielding superior performance on multiple code generation benchmarks. Additionally, curriculum learning based on CodeRM-NT further enhances model performance. Our code and dataset are available at: https://github.com/THUDM/CodeRM-NT.

Beining Wu, Fuyou Mao, Jiong Lin, Cheng Yang, Jiaxuan Lu, Yifu Guo, Siyu Zhang, Yifan Wu, Ying Huang, Fu Li

Generative engines (GEs) are reshaping information access by replacing ranked links with citation-grounded answers, yet current Generative Engine Optimization (GEO) methods optimize each instance in isolation, unable to accumulate or transfer effective strategies across tasks and engines. We reframe GEO as a strategy learning problem and propose MAGEO, a multi-agent framework in which coordinated planning, editing, and fidelity-aware evaluation serve as the execution layer, while validated editing patterns are progressively distilled into reusable, engine-specific optimization skills. To enable controlled assessment, we introduce a Twin Branch Evaluation Protocol for causal attribution of content edits and DSV-CF, a dual-axis metric that unifies semantic visibility with attribution accuracy. We further release MSME-GEO-Bench, a multi-scenario, multi-engine benchmark grounded in real-world queries. Experiments on three mainstream engines show that MAGEO substantially outperforms heuristic baselines in both visibility and citation fidelity, with ablations confirming that engine-specific preference modeling and strategy reuse are central to these gains, suggesting a scalable learning-driven paradigm for trustworthy GEO. Code is available at https://github.com/Wu-beining/MAGEO.

Gabeen Kim, Kyeongpil Kang

Historical documents act as invaluable knowledge archives but often suffer from illegibility due to physical deterioration and damage. While existing restoration methods based on masked language modeling effectively utilize local context, they struggle to restore named entities that require external historical knowledge. To address this limitation, we introduce a novel framework for historical document restoration that leverages large language models with retrieval-augmented generation (RAG). By combining the implicit knowledge of pre-trained LLMs with explicitly retrieved external context, our model ARI effectively mitigates the challenge of inferring context-dependent proper nouns. Extensive experiments on Korean historical documents demonstrate that our approach significantly outperforms baselines, achieving substantial gains in restoring both general characters and named entities. Furthermore, comprehensive evaluations including expert assessments confirm that ARI serves as a practical tool for domain experts, promising to accelerate the analysis of historical records.

Jonas Waldendorf, Bashar Awwad Shiekh Hasan, Evgenii Tsymbalov

Hallucinations in Speech Large Language Models (SpeechLLMs) pose significant risks, yet existing detection methods typically rely on gold-standard outputs that are costly or impractical to obtain. Moreover, hallucination detection methods developed for text-based LLMs do not directly capture audio-specific signals. We investigate four attention-derived metrics: AudioRatio, AudioConsistency, AudioEntropy, and TextEntropy, designed to capture pathological attention patterns associated with hallucination, and train lightweight logistic regression classifiers on these features for efficient inference-time detection. Across automatic speech recognition and speech-to-text translation tasks, evaluations on Qwen-2-Audio and Voxtral-3B show that our approach outperforms uncertainty-based and prior attention-based baselines on in-domain data, achieving improvements of up to +0.23 PR-AUC, and generalises to out-of-domain ASR settings. We further find that strong performance can be achieved with approximately 100 attention heads, improving out-of-domain generalisation compared to using all heads. While effectiveness is model-dependent and task-specific training is required, our results demonstrate that attention patterns provide a valuable tool for hallucination detection in SpeechLLMs

Hojae Han, Yeonseok Jeong, Seung-won Hwang, Zhewei Yao, Yuxiong He

Modern Text-to-SQL systems generate multiple candidate SQL queries and rank them to judge a final prediction. However, existing methods face two limitations. First, they often score functionally equivalent SQL queries inconsistently despite identical execution results. Second, ranking cannot recover when the correct SQL is absent from the candidate pool. We propose R3 -SQL, a Text-to-SQL framework that addresses both issues through unified reward for ranking and resampling. R3 -SQL first groups candidates by execution result and ranks groups for consistency. To score each group, it combines a pairwise preference across groups with a pointwise utility from the best group rank and size, capturing relative preference, consistency, and candidate quality. To improve candidate recall, R3 -SQL introduces agentic resampling, which judges the generated candidate pool and selectively resamples when the correct SQL is likely absent. R3 -SQL achieves 75.03 execution accuracy on BIRD-dev, a new state of the art among methods using models with disclosed sizes, with consistent gains across five benchmarks.

Maximo Eduardo Rulli, Simone Petruzzi, Edoardo Michielon, Fabrizio Silvestri, Simone Scardapane, Alessio Devoto

Masked Diffusion Language Models (DLMs) have recently emerged as a promising alternative to traditional Autoregressive Models (ARMs). DLMs employ transformer encoders with bidirectional attention, enabling parallel token generation while maintaining competitive performance. Although their efficiency and effectiveness have been extensively studied, the internal mechanisms that govern DLMs remain largely unexplored. In this work, we conduct an empirical analysis of DLM attention patterns, focusing on the attention sinking phenomenon, an effect previously observed in various transformer-based architectures. Our findings reveal that DLMs also exhibit attention sinks, but with distinct characteristics. First, unlike in ARMs, the sink positions in DLMs tend to shift throughout the generation process, displaying a dynamic behaviour. Second, while ARMs are highly sensitive to the removal of attention sinks, DLMs remain robust: masking sinks leads to only a minor degradation in performance. These results provide new insights into the inner workings of diffusion-based language models and highlight fundamental differences in how they allocate and utilize attention compared to autoregressive models.

Hojae Han, Jaejin Kim, Seung-won Hwang, Yu Jin Kim, Moontae Lee

This work addresses test output prediction, a key challenge in test case generation. To improve the reliability of predicted outputs by LLMs, prior approaches generate code first to ground predictions. One grounding strategy is direct execution of generated code, but even minor errors can cause failures. To address this, we introduce LLM-based pseudocode execution, which grounds prediction on more error-resilient pseudocode and simulates execution via LLM reasoning. We further propose DUET, a dual-execution framework that combines both approaches by functional majority voting. Our analysis shows the two approaches are complementary in overcoming the limitations of direct execution suffering from code errors, and pseudocode reasoning from hallucination. On LiveCodeBench, DUET achieves the state-of-the-art performance, improving Pass@1 by 13.6 pp. For filtering candidates in code generation, DUET shows the best Pass@1 on LiveCodeBenchEasy, BigCodeBench-Hard, DevEval and HumanEval(+).