论文检索

Siyi Li, Jiajun Shi, Shiwen Ni, Ge Zhang, Shuaimin Li, Shijian Wang, Zhoufutu Wen, Yizhi LI, Hamid Alinejad-Rokny, Jiaheng Liu 等

Large Reasoning Models (LRMs) have demonstrated strong performance by producing extended Chain-of-Thought (CoT) traces before answering. However, this paradigm often induces over-reasoning: redundant calculations and circular self-verification that increase computational cost without improving outcomes. Existing evaluations largely emphasize final accuracy or coarse token counts, and lack automated tools to separate essential logic from structural redundancy. We introduce CoTJudger, a graph-driven framework that quantifies reasoning efficiency by converting free-form CoTs into directed dependency graphs and extracting the Shortest Effective Path (SEP) needed to reach a correct solution. This yields an interpretable efficiency signal – how much of a CoT is necessary versus structurally redundant – that is comparable across models and tasks. Evaluating 21 LRMs, CoTJudger reveals pervasive redundancy and surfaces recurring failure modes, including verification obsession and compensatory redundancy. These results provide a practical metric for disentangling reasoning ability from computational waste, enabling more targeted evaluation and diagnosis of LRM efficiency.

Yi-Li Hsu, Li-Wun Chang, Chun-Yu Hsu, Wei-Kuan Shih, Aiping Xiong, Lun-Wei Ku

As AI systems increasingly mediate everyday communication, large language models (LLMs) are expected not only to provide factually accurate responses but also to generate explanations that engage with users’ mental states. We build on the concept of cognitive chains—structured representations of Situation, Clue, Thought, Action, and Emotion inspired by Theory of Mind—to investigate whether conditioning LLM outputs on such belief chains improves explanation quality. Specifically, we evaluate explanations along six reader-perceived dimensions: overall quality, logical correctness, completeness, conciseness, empathy, and agreement. Prior work shows that LLM explanations often default to neutral or uncertain stances, while individuals holding strong false beliefs remain highly resistant to correction. To address this challenge, we instantiate cognitive chains from two perspectives: believers and non-believers of the news claims. Using GPT-4.1 as a role-player across these stances, we find that incorporating believers’ chains improves the perceived quality of explanations for audiences with misinformation-aligned beliefs. Our findings underscore the importance of modeling diverse mental states in explanation generation and provide the first systematic evidence that Theory-of-Mind–based cognitive chains enhance the persuasiveness of explanations in misinformation contexts.

Gejian Zhao, Hanzhou Wu, Xinpeng Zhang

Large language models (LLMs) are increasingly expected to support pluralistic alignment, representing diverse human perspectives. However, current methods often induce motivated reasoning: LLMs tend to hallucinate “convenient” facts to forcefully justify a requested stance. To address this, we propose Value-Graph-Consistent Chain-of-Thought (VGC-CoT), a neuro-symbolic framework that enables steerable pluralism without distorting objective reality. We enforce a strict distinction: facts should be shared, while value trade-offs may diverge. Our approach models reasoning as a directed traversal over a multi-perspective graph comprising a fixed factual layer and perspective-specific value layers. By projecting generated CoT paths onto this structure, we align the model with target values while constraining it to a shared factual backbone. Experiments show that our method reduces factual hallucinations by 3\times and improves cross-perspective consistency by 25% compared to standard steerable baselines, paving the way for trustworthy pluralistic AI.

Fan Huang, Haewoon Kwak, Jisun An

Large Language Models (LLMs) are increasingly employed in various question-answering tasks. However, recent studies showcase that LLMs are susceptible to persuasion and could adopt counterfactual beliefs.We present a systematic evaluation of LLM susceptibility to persuasion under the Source–Message–Channel–Receiver (SMCR) communication framework. Across six mainstream Large Language Models (LLMs) and three domains (factual knowledge, medical QA, and social bias), we analyze how different persuasive strategies influence stated belief stability over multiple interaction turns.We further examine whether verbalized confidence prompting (i.e., eliciting self-reported confidence scores) affects resistance to persuasion.Results show that the smallest model (Llama 3.2-3B) exhibits extreme compliance, with 82.5% of belief changes occurring at the first persuasive turn (average end turn of 1.1–1.4).Contrary to expectations, verbalized confidence prompting increases vulnerability by accelerating belief erosion rather than enhancing robustness. Finally, an exploratory study of adversarial fine-tuning reveals highly model-dependent effectiveness: GPT-4o-mini achieves near-complete robustness (98.6%) and Mistral 7B improves substantially (35.7% \rightarrow 79.3%), but Llama models remain highly susceptible (<14% RQ1) even when fine-tuned on their own failure cases. Together, these findings highlight substantial model-dependent limits of current robustness interventions and offer guidance for developing more trustworthy LLMs[<https://github.com/muyuhuatang/llm_stated_belief>].

John Chen, Alexandros Nikolaos Lotsos, Sihan Cheng, Lexie Zhao, Yanjia Zhang, Jessica Hullman, Bruce Sherin, Uri Wilensky, Michael Horn

Qualitative analysis is critical to understanding human datasets in many social science disciplines. A central method in this process is inductive coding, where researchers identify and interpret codes directly from the datasets themselves. Yet, this exploratory approach poses challenges for meeting methodological expectations (such as "depth" and "variation"), especially as researchers increasingly adopt Generative AI (GAI) for support. Ground-truth-based metrics are insufficient because they contradict the exploratory nature of inductive coding; cluster- or topic-level metrics fail to capture the interpretive, cross-cutting nature of qualitative codes; and manual evaluation can be labor-intensive. This paper presents a theory-informed computational method for measuring inductive coding results from humans and GAI. Our method first merges individual codebooks into an Aggregated Code Space using an LLM-enriched hierarchical clustering algorithm. It then measures each coder’s contribution against the merged result using four novel metrics: Coverage, Overlap, Novelty, and Divergence, designed to capture breadth, consensus, unique contribution, and systematic deviation without assuming ground truth. Through two experiments on a human-coded online conversation dataset, we 1) reveal the merging algorithm’s impact on metrics; 2) validate the metrics’ stability and robustness across multiple runs and different LLMs; and 3) showcase the metrics’ ability to diagnose coding issues, such as excessive or irrelevant (hallucinated) codes. We discuss how these metrics should be interpreted in combination and their current limitations. Our work provides a reliable pathway for ensuring methodological rigor in human-AI qualitative analysis.

Ritam Upadhyay, Naman Ahuja, Rishabh Baral, Aparna Garimella, Vivek Gupta

Large language model (LLM) approaches to tabular summarization rely on extensive prompt engineering, decomposition pipelines, or entity-level intermediate representations to achieve strong performance. While effective, these strategies are computationally expensive and offer limited insight into how well models maintain state over long, evolving narratives. We introduce SporTabSet, a diagnostic benchmark for long-context tabular summarization across two complementary sports domains that require tracking multiple entities and aggregating statistics under domain-specific rules. Using SporTabSet, we systematically evaluate decomposition-based strategies across several long context LLMs. Results show that although decomposition substantially improves accuracy and numerical fidelity, gains stem mainly from dissecting multi-entity interference rather than improved local arithmetic. Robustness experiments further reveal high sensitivity to surface-level cues with structured failures, including hallucination, omission, and role confusion. Together, these findings identify consistent multi-entity memory as a key bottleneck in long-context table generation, motivating diagnostic evaluation as a prerequisite for scalable, efficient, and reliable tabular summarization models.

Riza Setiawan Soetedjo, Yusuke Sakai, Hidetaka Kamigaito, Jingun Kwon, Manabu Okumura, Taro Watanabe

Improving the quality of model-generated summaries, especially factuality, the accuracy of a summary with respect to its source content, remains a challenge. While reranking could select the optimal output from multiple generated candidates, it is limited to only using the source as guidance, resulting in unreliable summaries. To address this limitation, we propose ConSUM that reranks candidate summaries by considering two factors: consistency to the source document and consensus among the other candidates. Consensus is established using Minimum Bayes Risk (MBR) decoding over the set of generated summaries, while ensuring consistency by employing factuality-aware metrics that compare the summary against the source. Rigorous testing demonstrates that our system is competitive with existing methods, with human evaluations further confirming that its generated summaries are preferred over those from other systems.

Neeva Oza, Ishaan Govil, Parul Gupta, Dinesh Khandelwal, Dinesh Garg, Parag Singla

We study how well LLMs can determine whether two programs are functionally equivalent. This is an important problem because benchmarking code equivalence helps assess LLM capability in tasks such as code rewriting and translation. To this end, we introduce CETBench — Code Equivalence with Transformations Benchmark — built from a repository of programs that may solve the same or different tasks. Each dataset instance is created by sampling a program pair and applying a random sequence of predefined code transformations, yielding either equivalent or non-equivalent pairs. Our analysis shows that even simple transformations cause a significant performance drop in state-of-the-art LLMs on code-equivalence checking. These challenges are further amplified in the cross-lingual setting when comparing programs written in different languages. To remedy this, we present a simple fine-tuning-based approach to boost LLM performance on the transformed pairs of programs. Our approach for dataset generation is generic, supporting cross-lingual equivalence checking, the generation of program pairs with varying difficulty levels, and the application of diverse transformations. In our experiments, we perform ablations over the difficulty level of original programs, as well as the kind of transformations used in generating pairs for equivalence checking. Our analysis presents deep insights into the working of LLMs for the task of code-equivalence, and points to the fact that they may still be far from what could be termed as a semantic understanding of the underlying code.

Jinghao Luo, Yuchen Tian, Chuxue Cao, Ziyang Luo, Hongzhan Lin, Kaixin Li, Chuyi Kong, Ruichao Yang, Jing Ma

Large Language Model (LLM)-based agents have fundamentally reshaped artificial intelligence by integrating external tools and planning capabilities. While memory mechanisms have emerged as the architectural cornerstone of these systems, current research remains fragmented, oscillating between operating system engineering and cognitive science. This theoretical divide prevents a unified view of technological synthesis and a coherent evolutionary perspective. To bridge this gap, this survey proposes a novel evolutionary framework for LLM agent memory mechanisms, formalizing the development process into three stages: **Storage** (trajectory preservation), **Reflection** (trajectory refinement), and **Experience** (trajectory abstraction). We first formally define these three stages before analyzing the three core drivers of this evolution: the necessity for long-range consistency, the challenges in dynamic environments, and the ultimate goal of continual learning. Furthermore, we specifically explore two transformative mechanisms in the frontier Experience stage: proactive exploration and cross-trajectory abstraction. By synthesizing these disparate views, this work offers robust design principles and a clear roadmap for the development of next-generation LLM agents.

Ehsaneddin Asgari, Yassine El Kheir, MohammadAli SadraeiJavaheri, Ali Nazari

Tokenization fundamentally shapes NLP performance, affecting both efficiency and linguistic fidelity. While Byte Pair Encoding (BPE) underpins most Large Language Models (LLMs), its frequency-driven merges often disregard morpheme boundaries, yielding inconsistent and semantically opaque segmentations in morphologically rich languages. We introduce MorphBPE, a simple extension of BPE that constrains merge operations during tokenizer training to respect morpheme boundaries, while leaving inference unchanged and fully compatible with existing LLM pipelines. We evaluate tokenization quality using two intrinsic metrics, Morphological Consistency F1, which measures whether shared morphemes are assigned consistent token representations, and Morphological Edit Distance, which quantifies alignment with morpheme boundaries. We then train 300M and 1B parameter decoder-only LMs from scratch across four typologically diverse languages, English, Russian, Hungarian, and Arabic, under identical vocabulary sizes and training settings. Across all languages, MorphBPE consistently improves intrinsic morphological coherence and reduces language model cross-entropy, moreover, token length statistics indicate that these gains are not attributable to materially shorter tokens. Finally, on the Belebele multilingual reading comprehension benchmark, MorphBPE yields significant improvements in morphologically rich languages such as Russian and Arabic.

Hanna Yukhymenko, Anton Alexandrov, Martin Vechev

The reliability of multilingual Large Language Model (LLM) evaluation is currently compromised by the inconsistent quality of translated benchmarks. Existing resources often suffer from semantic drift and context loss, which can lead to misleading performance metrics. In this work, we present a fully automated framework designed to address these challenges by enabling scalable, high-quality translation of datasets and benchmarks. We demonstrate that adapting test-time compute scaling strategies, specifically Universal Self-Improvement (USI) and our proposed multi-round ranking method, T-RANK, allows for significantly higher quality outputs compared to traditional pipelines. Our framework ensures that benchmarks preserve their original task structure and linguistic nuances during localization. We apply this approach to translate popular benchmarks and datasets into eight Eastern and Southern European languages (Ukrainian, Bulgarian, Slovak, Romanian, Lithuanian, Estonian, Turkish, Greek). Evaluations using both reference-based metrics and LLM-as-a-judge show that our translations surpass existing resources, resulting in more accurate downstream model assessment. We release both the framework and the improved benchmarks to facilitate robust and reproducible multilingual AI development.

Shuo Han, Yukun Cao, Zezhong Ding, Zengyi Gao, S Kevin Zhou, Xike Xie

Vision-language models (VLMs) have shown promise in graph understanding, but remain limited by input-token constraints, facing scalability bottlenecks and lacking effective mechanisms to coordinate textual and visual modalities. To address these challenges, we propose GraphVista, a unified framework that enhances both scalability and modality coordination in graph understanding. For scalability, GraphVista organizes graph information hierarchically into a lightweight GraphRAG base, which retrieves only task-relevant textual descriptions and high-resolution visual subgraphs, compressing redundant context while preserving key reasoning elements. For modality coordination, GraphVista introduces a planning agent that routes tasks to the most suitable modality—using the text modality for simple property reasoning and the visual modality for local and structurally complex reasoning grounded in explicit topology. Extensive experiments demonstrate that GraphVista scales to large graphs, up to 200{\times} larger than those used in existing benchmarks, and consistently outperforms existing textual, visual, and fusion-based methods, achieving up to 4.4{\times} quality improvement over the state-of-the-art baselines by fully exploiting the complementary strengths of both modalities.

Guanyu Chen, Chenxiao Yu, Xiyang Hu

Large language models (LLMs) are increasingly used to simulate decision-making tasks involving personal data sharing, where privacy concerns and prosocial motivations can push choices in opposite directions. Existing evaluations often measure privacy-related attitudes or sharing intentions in isolation, which makes it difficult to determine whether a model’s expressed values jointly predict its downstream data-sharing actions as in real human behaviors. We introduce a context-based assessment protocol that sequentially administers standardized questionnaires for privacy attitudes, prosocialness, and acceptance of data sharing within a bounded, history-carrying session. To evaluate value-action alignments under competing attitudes, we use multi-group structural equation modeling (MGSEM) to identify relations from privacy concerns and prosocialness to data sharing. We propose Value-Action Alignment Rate (VAAR), a human-referenced directional agreement metric that aggregates path-level evidence for expected signs. Across multiple LLMs, we observe stable but model-specific Privacy-PSA-AoDS profiles, and substantial heterogeneity in value-action alignment.

Jiaxin Zhang, Wendi Cui, Zhuohang Li, Lifu Huang, Bradley A. Malin, Caiming Xiong, Chien-Sheng Wu

While Large Language Models (LLMs) show remarkable capabilities, their unreliability remains a critical barrier to deployment in high-stakes domains. This survey charts a functional evolution in addressing this challenge: the evolution of uncertainty from a passive diagnostic metric to an active control signal guiding real-time model behavior. We demonstrate how uncertainty is leveraged as an active control signal across three frontiers: in advanced reasoning to optimize computation and trigger self-correction; in autonomous agents to govern metacognitive decisions about tool use and information seeking; and in reinforcement learning to mitigate reward hacking and enable self-improvement via intrinsic rewards. By grounding these advancements in emerging theoretical frameworks like Bayesian methods and Conformal Prediction, we provide a unified perspective on this transformative trend. This survey provides a comprehensive overview, critical analysis, and practical design patterns, arguing that mastering the new trend of uncertainty is essential for building the next generation of scalable, reliable, and trustworthy AI.

Jie Zeng, Yukiko Nakano

The primary goal of Motivational Interviewing (MI) is to help clients build their own motivation for behavioral change. To support this in dialogue systems, it is essential to guide large language models (LLMs) to generate counselor responses aligned with MI principles. By employing a schema-guided approach, this study proposes a method for updating multi-frame dialogue states and a strategy decision mechanism that dynamically determines the response focus in a manner grounded in MI principles. The proposed method was implemented in a dialogue system on two different datasets and evaluated through a user study. Results showed that the proposed method successfully generates responses aligned with MI principle and frequently asks questions to elicit change talk.

Rishabh Kumar, Abhinav Painuli, Chriss Philip Saji, Devesh Soni, Amrith Krishna, Ganesh Ramakrishnan

Quality management when creating large-scale speech datasets is essential for building reliable downstream models, yet verification pipelines are often brittle, domain-specific, and expertise-intensive. We introduce SpeechQM-Agent, a natural language-driven agentic framework that compiles user requirements into dependency-aware DAG workflows over modular tools for audio, transcript, and metadata verification. A central planner LLM enforces prerequisites and supports execution-time replanning (e.g., re-running failed steps or swapping tools), reducing manual pipeline engineering and improving robustness across heterogeneous vendor formats and multilingual settings. We also release SpeechQM-Dataset, a multilingual benchmark with controlled, vendor-inspired quality artifacts spanning 24 verification tasks. Across experiments, SpeechQM-Agent attains 80-90% agreement with expert verification while requiring <20% of the cost and time of manual QC, and we further validate transfer to real vendor-supplied corpora. Planner LLM comparisons highlight fidelity-efficiency trade-offs.

Sneheel Sarangi, Hanan Salam

Recent advancements in large language models (LLMs) have demonstrated emergent capabilities in complex reasoning, largely spurred by rule-based Reinforcement Learning (RL) techniques applied during post-training. This has raised the question of whether similar methods can instill more nuanced, human-like social intelligence, such as a Theory of Mind (ToM), in LLMs. This paper investigates whether small- scale LLMs can acquire a robust and generalizable ToM capability through RL with verifiable rewards (RLVR). We conduct a systematic evaluation by training models on various combinations of prominent ToM benchmarks (HiToM, ExploreToM, FANToM) and testing for generalization on held-out benchmarks (e.g., Open- ToM). Our findings indicate that small LLMs struggle to develop a generic ToM capability. While performance on in-distribution tasks improves, this capability fails to transfer to unseen ToM tasks with different characteristics. Even observed out-of-distribution (OOD) performance improvements occur unpredictably across the training run, and don’t generalize across other OOD benchmarks. Furthermore, we conduct analysis to show that the learned behavior is likely a form of narrow overfitting rather than the acquisition of a true, abstract ToM capability.

Jonathan Ivey, Shivani Kumar, Jiayu Liu, Hua Shen, Sushrita Rakshit, Rohan Raju, Haotian Zhang, Aparna Ananthasubramaniam, Junghwan Kim, Bowen Yi 等

Building datasets for dialogue tasks is expensive and time-consuming, requiring recruitment, training, and data collection from study participants. In response, much recent work has sought to use large language models (LLMs) to simulate both human-human and human-LLM interactions, as they have been shown to generate convincingly human-like text in many settings. However, how well do LLM-based simulations reflect real human dialogue? In this work, we answer this question by generating a large-scale dataset of 100,000 paired LLM-LLM and human-LLM dialogues from the WildChat dataset and quantifying how well the LLM simulations align with their human counterparts. Overall, we find relatively low alignment between simulations and human interactions, with systematic differences in multiple textual properties, including style and conversational dynamics. Further, we find that models perform similarly in simulating English, Chinese, and Russian dialogues. Our results also suggest that LLMs only simulate a narrow range of the overall distribution of human dialogue, as they perform better on the subset of humans who write similarly to the LLM’s own style.

Hitoshi Ito, Naoto Shirai, Kazutaka Kinugawa, Hideya Mino, Rei Endo, Yoshihiko Kawai

Large language models (LLMs) have demonstrated strong performance in translation tasks. Subtitle translation presents unique challenges, such as preserving the original work’s worldview and the distinctive speaking styles of its characters. Achieving high-quality translations that reflect these stylistic nuances typically requires bilingual data for a specific movie, which is often scarce or unavailable. Thus, we propose a data augmentation method that uses LLMs to improve translation performance for specific movies, even when only a few hundred bilingual sentence pairs are available. The method expands source-side data by rewriting original subtitles using information that can be extracted from the context, such as character profiles and scene descriptions, to maintain the tone and thematic consistency of the movie. For translation, the augmented sentences are aligned with manually translated originals using structural similarity, which enables style-preserving bilingual data generation via one-shot learning. Experimental results show that data augmented using the proposed method effectively improves BLEU scores for film subtitle translation, and achieves superior stylistic quality in human evaluation.

Xinmiao Yu, Liwen Zhang, Xiaocheng Feng, Pengjun Xie, Jingren Zhou, Bing Qin, Yong Jiang

Large Language Model(LLM)-based agents have shown strong capabilities in web information seeking, with reinforcement learning (RL) becoming a key optimization paradigm. However, planning remains a bottleneck, as existing methods struggle with long-horizon strategies. Our analysis reveals a critical phenomenon—plan anchor—where the first reasoning step disproportionately impacts downstream behavior in long-horizon web reasoning tasks. Current RL algorithms, fail to account for this by uniformly distributing rewards across the trajectory.To address this, we propose Anchor-GRPO, a two-stage RL framework that decouples planning and execution. In Stage 1, the agent optimizes its first-step planning using fine-grained rubrics derived from self-play experiences and human calibration. In Stage 2, execution is aligned with the initial plan through sparse rewards, ensuring stable and efficient tool usage. We evaluate Anchor-GRPO on four benchmarks: BrowseComp, BrowseComp-Zh, GAIA, and XBench-DeepSearch. Across models from 3B to 30B, Anchor-GRPO outperforms baseline GRPO and First-step GRPO, improving task success and tool efficiency. Notably, WebAnchor-30B achieves 46.0% pass@1 on BrowseComp and 76.4% on GAIA. Anchor-GRPO also demonstrates strong scalability, getting higher accuracy as model size and context length increase.