PyTorchKR Papers

최신 논문

총 139,333편
태그·기간 등 상세 필터

카테고리 탐색

arXiv:2505.16722공식 코드 · Papers with Code

Breaking mBad! Supervised Fine-tuning for Cross-Lingual Detoxification

As large language models (LLMs) become increasingly prevalent in global applications, ensuring that they are toxicity-free across diverse linguistic contexts remains a critical challenge. We explore "Cross-lingual Detoxification", a cross-lingual paradigm that mitigates toxicity, enabling detoxification capabilities to transfer between high and low-resource languages across different script families. We analyze cross-lingual detoxification's effectiveness through 504 extensive settings to evaluate toxicity reduction in cross-distribution settings with limited data and investigate how mitigation impacts model performance on non-toxic tasks, revealing trade-offs between safety and knowledge preservation. Our code and dataset are publicly available at https://github.com/himanshubeniwal/Breaking-mBad.

arXiv:2505.16661공식 코드 · Papers with Code

A Japanese Language Model and Three New Evaluation Benchmarks for Pharmaceutical NLP

We present a Japanese domain-specific language model for the pharmaceutical field, developed through continual pretraining on 2 billion Japanese pharmaceutical tokens and 8 billion English biomedical tokens. To enable rigorous evaluation, we introduce three new benchmarks: YakugakuQA, based on national pharmacist licensing exams; NayoseQA, which tests cross-lingual synonym and terminology normalization; and SogoCheck, a novel task designed to assess consistency reasoning between paired statements. We evaluate our model against both open-source medical LLMs and commercial models, including GPT-4o. Results show that our domain-specific model outperforms existing open models and achieves competitive performance with commercial ones, particularly on terminology-heavy and knowledge-based tasks. Interestingly, even GPT-4o performs poorly on SogoCheck, suggesting that cross-sentence consistency reasoning remains an open challenge. Our benchmark suite offers a broader diagnostic lens for pharmaceutical NLP, covering factual recall, lexical variation, and logical consistency. This work demonstrates the feasibility of building practical, secure, and cost-effective language models for Japanese domain-specific applications, and provides reusable evaluation resources for future research in pharmaceutical and healthcare NLP. Our model, codes, and datasets are released at https://github.com/EQUES-Inc/pharma-LLM-eval.

arXiv:2505.16582공식 코드 · Papers with Code

O$^2$-Searcher: A Searching-based Agent Model for Open-Domain Open-Ended Question Answering

Large Language Models (LLMs), despite their advancements, are fundamentally limited by their static parametric knowledge, hindering performance on tasks requiring open-domain up-to-date information. While enabling LLMs to interact with external knowledge environments is a promising solution, current efforts primarily address closed-end problems. Open-ended questions, which characterized by lacking a standard answer or providing non-unique and diverse answers, remain underexplored. To bridge this gap, we present O$^2$-Searcher, a novel search agent leveraging reinforcement learning to effectively tackle both open-ended and closed-ended questions in the open domain. O$^2$-Searcher leverages an efficient, locally simulated search environment for dynamic knowledge acquisition, effectively decoupling the external world knowledge from model's sophisticated reasoning processes. It employs a unified training mechanism with meticulously designed reward functions, enabling the agent to identify problem types and adapt different answer generation strategies. Furthermore, to evaluate performance on complex open-ended tasks, we construct O$^2$-QA, a high-quality benchmark featuring 300 manually curated, multi-domain open-ended questions with associated web page caches. Extensive experiments show that O$^2$-Searcher, using only a 3B model, significantly surpasses leading LLM agents on O$^2$-QA. It also achieves SOTA results on various closed-ended QA benchmarks against similarly-sized models, while performing on par with much larger ones.

arXiv:2505.16576공식 코드 · Papers with Code

EMULATE: A Multi-Agent Framework for Determining the Veracity of Atomic Claims by Emulating Human Actions

Determining the veracity of atomic claims is an imperative component of many recently proposed fact-checking systems. Many approaches tackle this problem by first retrieving evidence by querying a search engine and then performing classification by providing the evidence set and atomic claim to a large language model, but this process deviates from what a human would do in order to perform the task. Recent work attempted to address this issue by proposing iterative evidence retrieval, allowing for evidence to be collected several times and only when necessary. Continuing along this line of research, we propose a novel claim verification system, called EMULATE, which is designed to better emulate human actions through the use of a multi-agent framework where each agent performs a small part of the larger task, such as ranking search results according to predefined criteria or evaluating webpage content. Extensive experiments on several benchmarks show clear improvements over prior work, demonstrating the efficacy of our new multi-agent framework.

arXiv:2505.16526공식 코드 · Papers with Code

EnSToM: Enhancing Dialogue Systems with Entropy-Scaled Steering Vectors for Topic Maintenance

Small large language models (sLLMs) offer the advantage of being lightweight and efficient, which makes them suitable for resource-constrained environments. However, sLLMs often struggle to maintain topic consistency in task-oriented dialogue systems, which is critical for scenarios such as service chatbots. Specifically, it is important to ensure that the model denies off-topic or malicious inputs and adheres to its intended functionality so as to prevent potential misuse and uphold reliability. Towards this, existing activation engineering approaches have been proposed to manipulate internal activations during inference. While these methods are effective in certain scenarios, our preliminary experiments reveal their limitations in ensuring topic adherence. Therefore, to address this, we propose a novel approach termed Entropy-scaled Steering vectors for Topic Maintenance (EnSToM). EnSToM dynamically adjusts the steering intensity based on input uncertainty, which allows the model to handle off-topic distractors effectively while preserving on-topic accuracy. Our experiments demonstrate that EnSToM achieves significant performance gain with a relatively small data size compared to fine-tuning approaches. By improving topic adherence without compromising efficiency, our approach provides a robust solution for enhancing sLLM-based dialogue systems.

arXiv:2505.16483공식 코드 · Papers with Code

Teaching Large Language Models to Maintain Contextual Faithfulness via Synthetic Tasks and Reinforcement Learning

Teaching large language models (LLMs) to be faithful in the provided context is crucial for building reliable information-seeking systems. Therefore, we propose a systematic framework, CANOE, to improve the faithfulness of LLMs in both short-form and long-form generation tasks without human annotations. Specifically, we first synthesize short-form question-answering (QA) data with four diverse tasks to construct high-quality and easily verifiable training data without human annotation. Also, we propose Dual-GRPO, a rule-based reinforcement learning method that includes three tailored rule-based rewards derived from synthesized short-form QA data, while simultaneously optimizing both short-form and long-form response generation. Notably, Dual-GRPO eliminates the need to manually label preference data to train reward models and avoids over-optimizing short-form generation when relying only on the synthesized short-form QA data. Experimental results show that CANOE greatly improves the faithfulness of LLMs across 11 different downstream tasks, even outperforming the most advanced LLMs, e.g., GPT-4o and OpenAI o1.

arXiv:2505.16467공식 코드 · Papers with Code

Reading Between the Prompts: How Stereotypes Shape LLM's Implicit Personalization

Generative Large Language Models (LLMs) infer user's demographic information from subtle cues in the conversation -- a phenomenon called implicit personalization. Prior work has shown that such inferences can lead to lower quality responses for users assumed to be from minority groups, even when no demographic information is explicitly provided. In this work, we systematically explore how LLMs respond to stereotypical cues using controlled synthetic conversations, by analyzing the models' latent user representations through both model internals and generated answers to targeted user questions. Our findings reveal that LLMs do infer demographic attributes based on these stereotypical signals, which for a number of groups even persists when the user explicitly identifies with a different demographic group. Finally, we show that this form of stereotype-driven implicit personalization can be effectively mitigated by intervening on the model's internal representations using a trained linear probe to steer them toward the explicitly stated identity. Our results highlight the need for greater transparency and control in how LLMs represent user identity.

arXiv:2505.16349공식 코드 · Papers with Code

Ask, Retrieve, Summarize: A Modular Pipeline for Scientific Literature Summarization

The exponential growth of scientific publications has made it increasingly difficult for researchers to stay updated and synthesize knowledge effectively. This paper presents XSum, a modular pipeline for multi-document summarization (MDS) in the scientific domain using Retrieval-Augmented Generation (RAG). The pipeline includes two core components: a question-generation module and an editor module. The question-generation module dynamically generates questions adapted to the input papers, ensuring the retrieval of relevant and accurate information. The editor module synthesizes the retrieved content into coherent and well-structured summaries that adhere to academic standards for proper citation. Evaluated on the SurveySum dataset, XSum demonstrates strong performance, achieving considerable improvements in metrics such as CheckEval, G-Eval and Ref-F1 compared to existing approaches. This work provides a transparent, adaptable framework for scientific summarization with potential applications in a wide range of domains. Code available at https://github.com/webis-de/scolia25-xsum

arXiv:2505.16234공식 코드 · Papers with Code

LIFEBench: Evaluating Length Instruction Following in Large Language Models

While large language models (LLMs) can solve PhD-level reasoning problems over long context inputs, they still struggle with a seemingly simpler task: following explicit length instructions-e.g., write a 10,000-word novel. Additionally, models often generate far too short outputs, terminate prematurely, or even refuse the request. Existing benchmarks focus primarily on evaluating generations quality, but often overlook whether the generations meet length constraints. To this end, we introduce Length Instruction Following Evaluation Benchmark (LIFEBench) to comprehensively evaluate LLMs' ability to follow length instructions across diverse tasks and a wide range of specified lengths. LIFEBench consists of 10,800 instances across 4 task categories in both English and Chinese, covering length constraints ranging from 16 to 8192 words. We evaluate 26 widely-used LLMs and find that most models reasonably follow short-length instructions but deteriorate sharply beyond a certain threshold. Surprisingly, almost all models fail to reach the vendor-claimed maximum output lengths in practice, as further confirmed by our evaluations extending up to 32K words. Even long-context LLMs, despite their extended input-output windows, counterintuitively fail to improve length-instructions following. Notably, Reasoning LLMs outperform even specialized long-text generation models, achieving state-of-the-art length following. Overall, LIFEBench uncovers fundamental limitations in current LLMs' length instructions following ability, offering critical insights for future progress.

arXiv:2505.16232공식 코드 · Papers with Code

MuseRAG: Idea Originality Scoring At Scale

An objective, face-valid way to assess the originality of creative ideas is to measure how rare each idea is within a population -- an approach long used in creativity research but difficult to automate at scale. Tabulating response frequencies via manual bucketing of idea rephrasings is labor-intensive, error-prone, and brittle under large corpora. We introduce a fully automated, psychometrically validated pipeline for frequency-based originality scoring. Our method, MuseRAG, combines large language models (LLMs) with an externally orchestrated retrieval-augmented generation (RAG) framework. Given a new idea, the system retrieves semantically similar prior idea buckets and zero-shot prompts the LLM to judge whether the new idea belongs to an existing bucket or forms a new one. The resulting buckets enable computation of frequency-based originality metrics. Across five datasets (N=1143, n_ideas=16294), MuseRAG matches human annotators in idea clustering structure and resolution (AMI = 0.59) and in participant-level scoring (r = 0.89) -- while exhibiting strong convergent and external validity. Our work enables intent-sensitive, human-aligned originality scoring at scale to aid creativity research.

arXiv:2505.16170공식 코드 · Papers with Code

When Do LLMs Admit Their Mistakes? Understanding the Role of Model Belief in Retraction

Can large language models (LLMs) admit their mistakes when they should know better? In this work, we define the behavior of acknowledging errors in previously generated answers as "retraction" and aim to understand when and why LLMs choose to retract. We first construct model-specific datasets to evaluate whether a model will retract an incorrect answer that contradicts its own parametric knowledge. While LLMs are capable of retraction, they do so only infrequently. We demonstrate that retraction is closely tied to previously identified indicators of models' internal belief: models fail to retract wrong answers that they "believe" to be factually correct. Steering experiments further demonstrate that internal belief causally influences model retraction. In particular, when the model does not believe its answer, this not only encourages the model to attempt to verify the answer, but also alters attention behavior during self-verification. Finally, we demonstrate that simple supervised fine-tuning significantly improves retraction performance by helping the model learn more accurate internal beliefs. Code and datasets are available on https://github.com/ayyyq/llm-retraction.

arXiv:2505.16160공식 코드 · Papers with Code

EduBench: A Comprehensive Benchmarking Dataset for Evaluating Large Language Models in Diverse Educational Scenarios

As large language models continue to advance, their application in educational contexts remains underexplored and under-optimized. In this paper, we address this gap by introducing the first diverse benchmark tailored for educational scenarios, incorporating synthetic data containing 9 major scenarios and over 4,000 distinct educational contexts. To enable comprehensive assessment, we propose a set of multi-dimensional evaluation metrics that cover 12 critical aspects relevant to both teachers and students. We further apply human annotation to ensure the effectiveness of the model-generated evaluation responses. Additionally, we succeed to train a relatively small-scale model on our constructed dataset and demonstrate that it can achieve performance comparable to state-of-the-art large models (e.g., Deepseek V3, Qwen Max) on the test set. Overall, this work provides a practical foundation for the development and evaluation of education-oriented language models. Code and data are released at https://github.com/ybai-nlp/EduBench.

arXiv:2505.16129공식 코드 · Papers with Code

LLMs Are Not Scorers: Rethinking MT Evaluation with Generation-Based Methods

Recent studies have applied large language models (LLMs) to machine translation quality estimation (MTQE) by prompting models to assign numeric scores. Nonetheless, these direct scoring methods tend to show low segment-level correlation with human judgments. In this paper, we propose a generation-based evaluation paradigm that leverages decoder-only LLMs to produce high-quality references, followed by semantic similarity scoring using sentence embeddings. We conduct the most extensive evaluation to date in MTQE, covering 8 LLMs and 8 language pairs. Empirical results show that our method outperforms both intra-LLM direct scoring baselines and external non-LLM reference-free metrics from MTME. These findings demonstrate the strength of generation-based evaluation and support a shift toward hybrid approaches that combine fluent generation with accurate semantic assessment.

arXiv:2505.16107공식 코드 · Papers with Code

MPL: Multiple Programming Languages with Large Language Models for Information Extraction

Recent research in information extraction (IE) focuses on utilizing code-style inputs to enhance structured output generation. The intuition behind this is that the programming languages (PLs) inherently exhibit greater structural organization than natural languages (NLs). This structural advantage makes PLs particularly suited for IE tasks. Nevertheless, existing research primarily focuses on Python for code-style simulation, overlooking the potential of other widely-used PLs (e.g., C++ and Java) during the supervised fine-tuning (SFT) phase. In this research, we propose \textbf{M}ultiple \textbf{P}rogramming \textbf{L}anguages with large language models for information extraction (abbreviated as \textbf{MPL}), a novel framework that explores the potential of incorporating different PLs in the SFT phase. Additionally, we introduce \texttt{function-prompt} with virtual running to simulate code-style inputs more effectively and efficiently. Experimental results on a wide range of datasets demonstrate the effectiveness of MPL. Furthermore, we conduct extensive experiments to provide a comprehensive analysis. We have released our code for future research.

arXiv:2505.16036공식 코드 · Papers with Code

OpenEthics: A Comprehensive Ethical Evaluation of Open-Source Generative Large Language Models

Generative large language models present significant potential but also raise critical ethical concerns. Most studies focus on narrow ethical dimensions, and also limited diversity of languages and models. To address these gaps, we conduct a broad ethical evaluation of 29 recent open-source large language models using a novel data collection including four ethical aspects: Robustness, reliability, safety, and fairness. We analyze model behavior in both a commonly used language, English, and a low-resource language, Turkish. Our aim is to provide a comprehensive ethical assessment and guide safer model development by filling existing gaps in evaluation breadth, language coverage, and model diversity. Our experimental results, based on LLM-as-a-Judge, reveal that optimization efforts for many open-source models appear to have prioritized safety and fairness, and demonstrated good robustness while reliability remains a concern. We demonstrate that ethical evaluation can be effectively conducted independently of the language used. In addition, models with larger parameter counts tend to exhibit better ethical performance, with Gemma and Qwen models demonstrating the most ethical behavior among those evaluated.

arXiv:2505.15960공식 코드 · Papers with Code

Training Step-Level Reasoning Verifiers with Formal Verification Tools

Process Reward Models (PRMs), which provide step-by-step feedback on the reasoning generated by Large Language Models (LLMs), are receiving increasing attention. However, two key research gaps remain: collecting accurate step-level error labels for training typically requires costly human annotation, and existing PRMs are limited to math reasoning problems. In response to these gaps, this paper aims to address the challenges of automatic dataset creation and the generalization of PRMs to diverse reasoning tasks. To achieve this goal, we propose FoVer, an approach for training PRMs on step-level error labels automatically annotated by formal verification tools, such as Z3 for formal logic and Isabelle for theorem proof, which provide automatic and accurate verification for symbolic tasks. Using this approach, we synthesize a training dataset with error labels on LLM responses for formal logic and theorem proof tasks without human annotation. Although this data synthesis is feasible only for tasks compatible with formal verification, we observe that LLM-based PRMs trained on our dataset exhibit cross-task generalization, improving verification across diverse reasoning tasks. Specifically, PRMs trained with FoVer significantly outperform baseline PRMs based on the original LLMs and achieve competitive or superior results compared to state-of-the-art PRMs trained on labels annotated by humans or stronger models, as measured by step-level verification on ProcessBench and Best-of-K performance across 12 reasoning benchmarks, including MATH, AIME, ANLI, MMLU, and BBH. The datasets, models, and code are provided at https://github.com/psunlpgroup/FoVer.

arXiv:2503.18968공식 코드 · Papers with Code

MedAgent-Pro: Towards Evidence-based Multi-modal Medical Diagnosis via Reasoning Agentic Workflow

In modern medicine, clinical diagnosis relies on the comprehensive analysis of primarily textual and visual data, drawing on medical expertise to ensure systematic and rigorous reasoning. Recent advances in large Vision-Language Models (VLMs) and agent-based methods hold great potential for medical diagnosis, thanks to the ability to effectively integrate multi-modal patient data. However, they often provide direct answers and draw empirical-driven conclusions without quantitative analysis, which reduces their reliability and clinical usability. We propose MedAgent-Pro, a new agentic reasoning paradigm that follows the diagnosis principle in modern medicine, to decouple the process into sequential components for step-by-step, evidence-based reasoning. Our MedAgent-Pro workflow presents a hierarchical diagnostic structure to mirror this principle, consisting of disease-level standardized plan generation and patient-level personalized step-by-step reasoning. To support disease-level planning, an RAG-based agent is designed to retrieve medical guidelines to ensure alignment with clinical standards. For patient-level reasoning, we propose to integrate professional tools such as visual models to enable quantitative assessments. Meanwhile, we propose to verify the reliability of each step to achieve evidence-based diagnosis, enforcing rigorous logical reasoning and a well-founded conclusion. Extensive experiments across a wide range of anatomical regions, imaging modalities, and diseases demonstrate the superiority of MedAgent-Pro to mainstream VLMs, agentic systems and state-of-the-art expert models. Ablation studies and human evaluation by clinical experts further validate its robustness and clinical relevance. Code is available at https://github.com/jinlab-imvr/MedAgent-Pro.

arXiv:2505.16944공식 코드 · Papers with Code

AGENTIF: Benchmarking Instruction Following of Large Language Models in Agentic Scenarios

Large Language Models (LLMs) have demonstrated advanced capabilities in real-world agentic applications. Growing research efforts aim to develop LLM-based agents to address practical demands, introducing a new challenge: agentic scenarios often involve lengthy instructions with complex constraints, such as extended system prompts and detailed tool specifications. While adherence to such instructions is crucial for agentic applications, whether LLMs can reliably follow them remains underexplored. In this paper, we introduce AgentIF, the first benchmark for systematically evaluating LLM instruction following ability in agentic scenarios. AgentIF features three key characteristics: (1) Realistic, constructed from 50 real-world agentic applications. (2) Long, averaging 1,723 words with a maximum of 15,630 words. (3) Complex, averaging 11.9 constraints per instruction, covering diverse constraint types, such as tool specifications and condition constraints. To construct AgentIF, we collect 707 human-annotated instructions across 50 agentic tasks from industrial application agents and open-source agentic systems. For each instruction, we annotate the associated constraints and corresponding evaluation metrics, including code-based evaluation, LLM-based evaluation, and hybrid code-LLM evaluation. We use AgentIF to systematically evaluate existing advanced LLMs. We observe that current models generally perform poorly, especially in handling complex constraint structures and tool specifications. We further conduct error analysis and analytical experiments on instruction length and meta constraints, providing some findings about the failure modes of existing LLMs. We have released the code and data to facilitate future research.

arXiv:2505.16826공식 코드 · Papers with Code

KTAE: A Model-Free Algorithm to Key-Tokens Advantage Estimation in Mathematical Reasoning

Recent advances have demonstrated that integrating reinforcement learning with rule-based rewards can significantly enhance the reasoning capabilities of large language models, even without supervised fine-tuning. However, prevalent reinforcement learning algorithms such as GRPO and its variants like DAPO, suffer from a coarse granularity issue when computing the advantage. Specifically, they compute rollout-level advantages that assign identical values to every token within a sequence, failing to capture token-specific contributions and hindering effective learning. To address this limitation, we propose Key-token Advantage Estimation (KTAE) - a novel algorithm that estimates fine-grained, token-level advantages without introducing additional models. KTAE leverages the correctness of sampled rollouts and applies statistical analysis to quantify the importance of individual tokens within a sequence to the final outcome. This quantified token-level importance is then combined with the rollout-level advantage to obtain a more fine-grained token-level advantage estimation. Empirical results show that models trained with GRPO+KTAE and DAPO+KTAE outperform baseline methods across five mathematical reasoning benchmarks. Notably, they achieve higher accuracy with shorter responses and even surpass R1-Distill-Qwen-1.5B using the same base model.

arXiv:2505.15859공식 코드 · Papers with Code

AutoData: A Multi-Agent System for Open Web Data Collection

The exponential growth of data-driven systems and AI technologies has intensified the demand for high-quality web-sourced datasets. While existing datasets have proven valuable, conventional web data collection approaches face significant limitations in terms of human effort and scalability. Current data-collecting solutions fall into two categories: wrapper-based methods that struggle with adaptability and reproducibility, and large language model (LLM)-based approaches that incur substantial computational and financial costs. To address these challenges, we propose AutoData, a novel multi-agent system for Automated web Data collection, that requires minimal human intervention, i.e., only necessitating a natural language instruction specifying the desired dataset. In addition, AutoData is designed with a robust multi-agent architecture, featuring a novel oriented message hypergraph coordinated by a central task manager, to efficiently organize agents across research and development squads. Besides, we introduce a novel hypergraph cache system to advance the multi-agent collaboration process that enables efficient automated data collection and mitigates the token cost issues prevalent in existing LLM-based systems. Moreover, we introduce Instruct2DS, a new benchmark dataset supporting live data collection from web sources across three domains: academic, finance, and sports. Comprehensive evaluations over Instruct2DS and three existing benchmark datasets demonstrate AutoData's superior performance compared to baseline methods. Case studies on challenging tasks such as picture book collection and paper extraction from surveys further validate its applicability. Our source code and dataset are available at https://github.com/GraphResearcher/AutoData.