ACL 2026 · 全部收录(459 篇)¶
仅列出通过预过滤且已打分的论文;分数是「评测方法学价值」,不是会议录用质量。 返回总览
| # | 论文 | 轨道 | 分数 | 链接 |
|---|---|---|---|---|
| 1 | BenchMarker: An Education-Inspired Toolkit for Highlighting Flaws in Multiple-Choice Benchmarks | Long | 9 | 页 · PDF |
| 2 | Beyond the Leaderboard: Rethinking Medical Benchmarks for Large Language Models | Long | 9 | 页 · PDF |
| 3 | BrowseComp-Plus: A Fair and Disentangled Evaluation Benchmark for Deep Search Agents | Long | 9 | 页 · PDF |
| 4 | Controllable Contamination Detection for Reliable LLM Evaluation with Statistical Guarantees | Long | 9 | 页 · PDF |
| 5 | Decentralized Arena: Towards Democratic and Scalable Automatic Evaluation of Language Models | Long | 9 | 页 · PDF |
| 6 | DeepFact: Co-Evolving Benchmarks and Agents for Deep Research Factuality | Long | 9 | 页 · PDF |
| 7 | Evaluating Legal Reasoning Traces with Legal Issue Tree Rubrics | Long | 9 | 页 · PDF |
| 8 | From Word Sequences to Behavioral Sequences: Adapting Modeling and Evaluation Paradigms for Longitudinal NLP | Long | 9 | 页 · PDF |
| 9 | GenPT: Beyond Self-Report for Reliable LLM Psychometrics via Generative Projective Testing | Long | 9 | 页 · PDF |
| 10 | Illusions of the Gold Standard: A Large-scale Analysis of Human Evaluation Protocols for Long-form Text Generation | Long | 9 | 页 · PDF |
| 11 | Label Effects: Shared Heuristic Reliance in Trust Assessment by Humans and LLM-as-a-Judge | Long | 9 | 页 · PDF |
| 12 | LLMEval-Fair: A Large-Scale Longitudinal Study on Robust and Fair Evaluation of Large Language Models | Long | 9 | 页 · PDF |
| 13 | Lost in Simulation: LLM-Simulated Users are Unreliable Proxies for Human Users in Agentic Evaluations | Long | 9 | 页 · PDF |
| 14 | Macaron: Controlled, Human-Written Benchmark for Multilingual and Multicultural Reasoning via Template-Filling | Long | 9 | 页 · PDF |
| 15 | MedEinst: Benchmarking the Einstellung Effect in Medical LLMs through Counterfactual Differential Diagnosis | Long | 9 | 页 · PDF |
| 16 | Mediocrity is the key for LLM as a Judge Anchor Selection | Long | 9 | 页 · PDF |
| 17 | MQM Re-Annotation: A Technique for Collaborative Evaluation of Machine Translation | Long | 9 | 页 · PDF |
| 18 | ROSE: An Intent-Centered Evaluation Metric for NL2SQL | Long | 9 | 页 · PDF |
| 19 | SPENCE: A Syntactic Probe for Detecting Contamination in NL2SQL Benchmarks | Long | 9 | 页 · PDF |
| 20 | Stress Testing Factual Consistency Metrics for Long-Document Summarization | Long | 9 | 页 · PDF |
| 21 | Test of Time: Rethinking Temporal Signal of Benchmark Contamination | Long | 9 | 页 · PDF |
| 22 | The Proxy Presumption: From Semantic Embeddings to Valid Social Measures | Long | 9 | 页 · PDF |
| 23 | User Perceptions vs. Proxy LLM Judges: Privacy and Helpfulness in LLM Responses to Privacy-Sensitive Scenarios | Long | 9 | 页 · PDF |
| 24 | VeriTaS: The First Dynamic Benchmark for Multimodal Automated Fact-Checking | Long | 9 | 页 · PDF |
| 25 | When Benchmarks Leak: Inference-Time Decontamination for LLMs | Long | 9 | 页 · PDF |
| 26 | Your Reasoning Benchmark May Not Test Reasoning: Revealing Perception Bottleneck in Abstract Reasoning Benchmarks | Long | 9 | 页 · PDF |
| 27 | A Functionality-Grounded Benchmark for Evaluating Web Agents in E-commerce Domains | Long | 8 | 页 · PDF |
| 28 | ACIArena: Toward Unified Evaluation for Agent Cascading Injection | Long | 8 | 页 · PDF |
| 29 | AdvancedIF: Rubric-Based Benchmarking and Reinforcement Learning for Advancing LLM Instruction Following | Long | 8 | 页 · PDF |
| 30 | AgentGym2: Benchmarking Large Language Model Agents in De-Idealized Real-World Environments | Long | 8 | 页 · PDF |
| 31 | Are We Using the Right Benchmark: An Evaluation Framework for Visual Token Compression Methods | Long | 8 | 页 · PDF |
| 32 | arXiv2Table: Toward Realistic Benchmarking and Evaluation for LLM-Based Literature-Review Table Generation | Long | 8 | 页 · PDF |
| 33 | AutoJudger: An Agent-Driven Framework for Efficient Benchmarking of MLLMs | Long | 8 | 页 · PDF |
| 34 | Benchmarking Deflection and Hallucination in Large Vision-Language Models | Long | 8 | 页 · PDF |
| 35 | Benchmarking Fine-Grained Error Detection in Multimodal Reasoning | Long | 8 | 页 · PDF |
| 36 | Benchmarking Web Agent Safety under E-commerce Deceptive Interfaces | Long | 8 | 页 · PDF |
| 37 | Beyond "I Don’t Know": Evaluating LLM Self-Awareness in Discriminating Data and Model Uncertainty | Long | 8 | 页 · PDF |
| 38 | Beyond Ranking: Fine-Grained Diagnostics and Self-Improvement for MLLMs | Long | 8 | 页 · PDF |
| 39 | Beyond Static Benchmarks: Synthesizing Harmful Content via Persona-based Simulation for Robust Evaluation | Long | 8 | 页 · PDF |
| 40 | Beyond the Last Frame: Process-aware Evaluation for Generative Video Reasoning | Long | 8 | 页 · PDF |
| 41 | Beyond Value Benchmarks: Measuring Value-Structure Alignment in Large Language Models via Symmetric Q-Sorts | Long | 8 | 页 · PDF |
| 42 | Bloom-Eval: A Hierarchical Evaluation Benchmark for Automatic Survey Generation Based on Bloom’s Taxonomy | Long | 8 | 页 · PDF |
| 43 | CAR-bench: Evaluating the Consistency and Limit-Awareness of LLM Agents under Real-World Uncertainty | Long | 8 | 页 · PDF |
| 44 | Chain-of-Thought as a Lens: Evaluating Structured Reasoning Alignment between Human Preferences and Large Language Models | Long | 8 | 页 · PDF |
| 45 | ChemReason-Bench: Benchmarking Large Language Models for Procedural Reasoning in Experimental Chemistry | Long | 8 | 页 · PDF |
| 46 | CIG: Measuring Conversational Information Gain in Deliberative Dialogues with Semantic Memory Dynamics | Long | 8 | 页 · PDF |
| 47 | CodeJudgeBench: Benchmarking LLM-as-a-Judge for Coding Tasks | Long | 8 | 页 · PDF |
| 48 | CogToM: A Comprehensive Theory of Mind Benchmark inspired by Human Cognition for Large Language Models | Long | 8 | 页 · PDF |
| 49 | CommonLID: Re-evaluating State-of-the-Art Language Identification Performance on Web Data | Long | 8 | 页 · PDF |
| 50 | Cross-Examination Framework: A Task-Agnostic Diagnostic for Information Fidelity in Text-to-Text Generation | Long | 8 | 页 · PDF |
| 51 | CxMP: A Linguistic Minimal-Pair Benchmark for Evaluating Constructional Understanding in Language Models | Long | 8 | 页 · PDF |
| 52 | DAC-Bench: A Decision-Aware Benchmark for Compositional Mobile GUI Tasks | Long | 8 | 页 · PDF |
| 53 | DeepPlanning: Benchmarking Long-Horizon Agentic Planning with Verifiable Constraints | Long | 8 | 页 · PDF |
| 54 | DR-Arena: an Automated Evaluation Framework for Deep Research Agents | Long | 8 | 页 · PDF |
| 55 | DREAM: Deep Research Evaluation with Agentic Metrics | Long | 8 | 页 · PDF |
| 56 | E2EDev: Benchmarking Large Language Models in End-to-End Software Development Task | Long | 8 | 页 · PDF |
| 57 | EmoS: A High-Fidelity Multimodal Benchmark for Fine-grained Streaming Emotional Understanding | Long | 8 | 页 · PDF |
| 58 | Evaluating Language Model Pluralism through In-the-wild Crowd Discussions | Long | 8 | 页 · PDF |
| 59 | Evaluation Pitfalls and Challenges in Multimedia Event Extraction | Long | 8 | 页 · PDF |
| 60 | EVM-QuestBench: An Execution-Grounded Benchmark for Natural-Language Transaction Code Generation | Long | 8 | 页 · PDF |
| 61 | FinChain: A Symbolic Benchmark for Verifiable Chain-of-Thought Financial Reasoning | Long | 8 | 页 · PDF |
| 62 | From Laboratory to Real-World Applications: Benchmarking Agentic Code Reasoning at the Repository Level | Long | 8 | 页 · PDF |
| 63 | Gap-K%: Measuring Top-1 Prediction Gap for Detecting Pretraining Data | Long | 8 | 页 · PDF |
| 64 | GeoRC: A Benchmark for Geolocation Reasoning Chains | Long | 8 | 页 · PDF |
| 65 | Gradient-Guided Multi-Judge Prompt Optimization | Long | 8 | 页 · PDF |
| 66 | Hard2Verify: A Step-Level Verification Benchmark for Open-Ended Frontier Math | Long | 8 | 页 · PDF |
| 67 | HoWToBench: Holistic Evaluation for LLM’s Capability in Human-level Writing using Tree of Writing | Long | 8 | 页 · PDF |
| 68 | HSCodeComp: A Realistic and Expert-level Agent Benchmark for Hierarchical Rule Application | Long | 8 | 页 · PDF |
| 69 | HypoEval: Hypothesis-Guided Evaluation for Natural Language Generation | Long | 8 | 页 · PDF |
| 70 | IF-CRITIC: Towards a Fine-Grained LLM Critic for Instruction-Following Evaluation | Long | 8 | 页 · PDF |
| 71 | IF-RewardBench: Benchmarking Judge Models for Instruction-Following Evaluation | Long | 8 | 页 · PDF |
| 72 | Inflated Excellence or True Performance? Rethinking Medical Diagnostic Benchmarks with Dynamic Evaluation | Long | 8 | 页 · PDF |
| 73 | Is EEG-to-Text Feasible in Real-World Scenarios? An In-Depth Analysis Using a Neuropsychology-Inspired Benchmark | Long | 8 | 页 · PDF |
| 74 | JurisBench: A Deep Benchmark for Assessing Large Language Models in Professional Legal Practice | Long | 8 | 页 · PDF |
| 75 | Knowledge without Wisdom: Measuring Misalignment between LLMs and Intended Impact | Long | 8 | 页 · PDF |
| 76 | League of LLMs: A Benchmark-Free Paradigm for Mutual Evaluation of Large Language Models | Long | 8 | 页 · PDF |
| 77 | LexRel: Benchmarking Legal Relation Extraction for Chinese Civil Cases | Long | 8 | 页 · PDF |
| 78 | LiveCLKTBench: Towards Reliable Evaluation of Cross-Lingual Knowledge Transfer in Multilingual LLMs | Long | 8 | 页 · PDF |
| 79 | LiveFact: A Dynamic, Time-Aware Benchmark for LLM-Driven Fake News Detection | Long | 8 | 页 · PDF |
| 80 | Locomo-Plus: Beyond-Factual Cognitive Memory Evaluation Framework for LLM Agents | Long | 8 | 页 · PDF |
| 81 | MADE: A Living Benchmark for Multi-Label Text Classification with Uncertainty Quantification of Medical Device Adverse Events | Long | 8 | 页 · PDF |
| 82 | MathSight: A Benchmark Exploring Have Vision-Language Models Really Seen in University-Level Mathematical Reasoning? | Long | 8 | 页 · PDF |
| 83 | Measuring Distribution Shift in User Prompts and Its Effects on LLM Performance | Long | 8 | 页 · PDF |
| 84 | Measuring Social Bias in Vision-Language Models with Face-Only Counterfactuals from Real Photos | Long | 8 | 页 · PDF |
| 85 | METER: Evaluating Multi-Level Contextual Causal Reasoning in Large Language Models | Long | 8 | 页 · PDF |
| 86 | MM-JudgeBias: A Benchmark for Evaluating Compositional Biases in MLLM-as-a-Judge | Long | 8 | 页 · PDF |
| 87 | MobileWorld: Benchmarking Autonomous Mobile Agents in Agent-User Interactive and MCP-Augmented Environments | Long | 8 | 页 · PDF |
| 88 | Monotonic Scaffolding as a Diagnostic Lens for Legal Reasoning in LLMs | Long | 8 | 页 · PDF |
| 89 | MPR-GUI: Benchmarking and Enhancing Multilingual Perception and Reasoning in GUI Agents | Long | 8 | 页 · PDF |
| 90 | MTR-Suite: A Framework for Evaluating and Synthesizing Conversational Retrieval Benchmarks | Long | 8 | 页 · PDF |
| 91 | Multi-Agent-as-Judge: Aligning LLM-Agent-Based Automated Evaluation with Multi-Dimensional Human Evaluation | Long | 8 | 页 · PDF |
| 92 | NaturalGAIA: A Verifiable Benchmark and Hierarchical Framework for Long-Horizon GUI Tasks | Long | 8 | 页 · PDF |
| 93 | Neo-Classic: A Benchmark for Evaluating Linguistic-Aesthetic Reasoning in Classical Chinese Poetry | Long | 8 | 页 · PDF |
| 94 | OctoBench: Benchmarking Scaffold-Aware Instruction Following in Repository-Grounded Agentic Coding | Long | 8 | 页 · PDF |
| 95 | Omni-RewardBench: Toward a Comprehensive Evaluation of Generative Reward Models Across Modalities | Long | 8 | 页 · PDF |
| 96 | PEAR: Pairwise Evaluation for Automatic Relative Scoring in Machine Translation | Long | 8 | 页 · PDF |
| 97 | Persona-Grounded Safety Evaluation of AI Companions in Multi-Turn Conversations | Long | 8 | 页 · PDF |
| 98 | PIArena: A Platform for Prompt Injection Evaluation | Long | 8 | 页 · PDF |
| 99 | PluRule: A Benchmark for Moderating Pluralistic Communities on Social Media | Long | 8 | 页 · PDF |
| 100 | PRBench: Large-Scale Expert Rubrics for Evaluating High-Stakes Professional Reasoning | Long | 8 | 页 · PDF |
| 101 | Putting HUMANS first: Efficient LAM Evaluation with Human Preference Alignment | Long | 8 | 页 · PDF |
| 102 | Quantifying Metric and Model Agreement in Bias Evaluation of Large Language Models | Long | 8 | 页 · PDF |
| 103 | Quantifying the Impact of Translation Errors on Multilingual LLM Evaluation | Long | 8 | 页 · PDF |
| 104 | Question Difficulty Estimation for Large Language Models via Answer Plausibility Scoring | Long | 8 | 页 · PDF |
| 105 | Reliable Evaluation Protocol for Low-Precision Retrieval | Short | 8 | 页 · PDF |
| 106 | ReproEvalCard: A Reporting Standard for Reproducible Evaluation of LLM Pipelines | Short | 8 | 页 · PDF |
| 107 | Rethinking Evaluation for LLM Hallucination Detection: A Desiderata, A New RAG-based Benchmark, New Insights | Long | 8 | 页 · PDF |
| 108 | ReTraceQA: Evaluating Reasoning Traces of Small Language Models in Commonsense Question Answering | Long | 8 | 页 · PDF |
| 109 | Revisiting Metric Reliability for Fine-grained Evaluation of Machine Translation and Summarization in Indian Languages | Long | 8 | 页 · PDF |
| 110 | Revisiting the Reliability of Language Models in Instruction-Following | Long | 8 | 页 · PDF |
| 111 | S2S-Arena: Evaluating Paralinguistic Instruction Following in Speech-to-Speech Models | Long | 8 | 页 · PDF |
| 112 | SAGE: A Search-AuGmented Evaluation of Large Language Models on Free-Form QA | Long | 8 | 页 · PDF |
| 113 | SATQuest: A Verifier for Logical Reasoning Evaluation and Reinforcement Fine-Tuning of LLMs | Long | 8 | 页 · PDF |
| 114 | SecureVibeBench: Benchmarking Secure Vibe Coding of AI Agents via Reconstructing Vulnerability-Introducing Scenarios | Long | 8 | 页 · PDF |
| 115 | Seeing the Whole Elephant: A Benchmark for Failure Attribution in LLM-based Multi-Agent Systems | Long | 8 | 页 · PDF |
| 116 | SILO-BENCH: A Scalable Environment for Evaluating Distributed Coordination in Multi-Agent LLM Systems | Long | 8 | 页 · PDF |
| 117 | Simple Agents, Biased Judges: Efficient Multi-Party Dialogue Generation & The Evaluation Gap | Long | 8 | 页 · PDF |
| 118 | SpeechLLM-as-Judges: Towards General and Interpretable Speech Quality Evaluation | Long | 8 | 页 · PDF |
| 119 | Still Between Us? Evaluating and Improving Voice Assistant Robustness to Third-Party Interruptions | Long | 8 | 页 · PDF |
| 120 | Subject-level Inference for Realistic Text Anonymization Evaluation | Long | 8 | 页 · PDF |
| 121 | SwissGov-RSD: A Human-annotated, Cross-lingual Benchmark for Token-level Recognition of Semantic Differences Between Related Documents | Long | 8 | 页 · PDF |
| 122 | TabReX: Tabular Referenceless eXplainable Evaluation | Long | 8 | 页 · PDF |
| 123 | The Visual Iconicity Challenge: Evaluating Vision-Language Models on Sign Language Form–Meaning Mapping | Long | 8 | 页 · PDF |
| 124 | Toward Robust Evaluation for Multilingual Grammatical Error Correction: Can Large Language Models Replace Human References? | Long | 8 | 页 · PDF |
| 125 | TRACE: Evaluating Execution Efficiency of LLM-Based Code Translation | Long | 8 | 页 · PDF |
| 126 | Translation or Recitation? Calibrating Evaluation Scores for Machine Translation of Extremely Low-Resource Languages | Short | 8 | 页 · PDF |
| 127 | Unlearners Can Lie: Evaluating and Improving Honesty in LLM Unlearning | Long | 8 | 页 · PDF |
| 128 | UrbanGeoEval: A City-Scale Benchmark for Evaluating Large Language Models in Geospatial Reasoning | Long | 8 | 页 · PDF |
| 129 | USB: A COMPREHENSIVE AND UNIFIED SAFETY EVALUATION BENCHMARK FOR MULTIMODAL LARGE LANGUAGE MODELS | Long | 8 | 页 · PDF |
| 130 | VC-Inspector: Advancing Reference-free Evaluation of Video Captions with Factual Analysis | Long | 8 | 页 · PDF |
| 131 | VisAidMath: Benchmarking Visual-Aided Mathematical Reasoning | Long | 8 | 页 · PDF |
| 132 | VL-Calibration: Decoupled Confidence Calibration for Large Vision-Language Models Reasoning | Long | 8 | 页 · PDF |
| 133 | When High Accuracy Hides Poor Calibration: Rethinking Confidence Evaluation in Transformer-Based Text Classification with Balanced Brier Score | Long | 8 | 页 · PDF |
| 134 | When KV Cache Reuse Fails in Multi-Agent Systems: Cross-Candidate Interaction is Crucial for LLM Judges | Long | 8 | 页 · PDF |
| 135 | When More Words Say Less: Decoupling Length and Specificity in Image Description Evaluation | Short | 8 | 页 · PDF |
| 136 | When Vision-Language Models Judge Without Seeing: Exposing Informativeness Bias | Long | 8 | 页 · PDF |
| 137 | YIELD: A Large-Scale Dataset and Evaluation Framework for Information Elicitation Agents | Long | 8 | 页 · PDF |
| 138 | "Newspaper Eat" Means "Not Tasty": A Taxonomy and Benchmark for Coded Language in Real-World Chinese Online Reviews | Long | 7 | 页 · PDF |
| 139 | A Multi-Agent Framework for Feature-Constrained Difficulty Control in Reading Comprehension Item Generation | Long | 7 | 页 · PDF |
| 140 | A Multilingual Social Bias Benchmark Incorporating Thinking Processes | Long | 7 | 页 · PDF |
| 141 | ADAPT: Benchmarking Commonsense Planning under Unspecified Affordance Constraints | Long | 7 | 页 · PDF |
| 142 | AEGIS: A Holistic Benchmark for Evaluating Forensic Analysis of AI-Generated Academic Images | Long | 7 | 页 · PDF |
| 143 | AgencyBench: Benchmarking the Frontiers of Autonomous Agents in 1M-Token Real-World Contexts | Long | 7 | 页 · PDF |
| 144 | Aligning Agents via Planning: A Benchmark for Trajectory-Level Reward Modeling | Long | 7 | 页 · PDF |
| 145 | All That Glisters Is Not Gold: A Benchmark for Reference-Free Counterfactual Financial Misinformation Detection | Long | 7 | 页 · PDF |
| 146 | Are LLMs Reliable Rankers? Rank Manipulation via Two-Stage Token Optimization | Long | 7 | 页 · PDF |
| 147 | Ascending the Infinite Ladder: Benchmarking Spatial Deformation Reasoning in Vision-Language Models | Long | 7 | 页 · PDF |
| 148 | Audio Jailbreak: An Open Comprehensive Benchmark for Jailbreaking Large Audio-Language Models | Long | 7 | 页 · PDF |
| 149 | Audio MultiChallenge: A Multi-Turn Evaluation of Spoken Dialogue Systems on Natural Human Interaction | Long | 7 | 页 · PDF |
| 150 | Automated Creativity Evaluation of Language Models Across Open-Ended Tasks | Long | 7 | 页 · PDF |