LLM 固定订阅页

《Four-Axis Decision Alignment for Long-Horizon Enterprise AI Agents》〔评测 / 应用 / 方法〕：Long-horizon enterprise agents make high-stakes decisions (loan underwriting, claims adjudication, clinical review, prior authorization)…

1. Four-Axis Decision Alignment for Long-Horizon Enterprise AI Agents · Score 162
   title matched "agent"；title matched "alignment"；summary matched "reasoning"
   原始来源
2. Cyber Defense Benchmark: Agentic Threat Hunting Evaluation for LLMs in SecOps · Score 149
   title matched "agent"；title matched "benchmark"；title matched "evaluation"
   原始来源
3. Taming Actor-Observer Asymmetry in Agents via Dialectical Alignment · Score 145
   title matched "agent"；title matched "alignment"；summary matched "reasoning"
   原始来源
4. Discovering a Shared Logical Subspace: Steering LLM Logical Reasoning via Alignment of Natural-Language and Symbolic Views · Score 130
   title matched "reasoning"；title matched "alignment"；summary matched "benchmark"
   原始来源
5. Revac: A Social Deduction Reasoning Agent · Score 127
   title matched "agent"；title matched "reasoning"；summary matched "evaluation"
   原始来源

《MathNet: a Global Multimodal Benchmark for Mathematical Reasoning and Retrieval》〔评测 / 数据 / 方法〕：Mathematical problem solving remains a challenging test of reasoning for large language and multimodal models, yet existing…

1. MathNet: a Global Multimodal Benchmark for Mathematical Reasoning and Retrieval · Score 112
   title matched "reasoning"；title matched "benchmark"；has PDF
   原始来源
2. Benchmarking System Dynamics AI Assistants: Cloud Versus Local LLMs on CLD Extraction and Discussion · Score 108
   title matched "benchmark"；summary matched "reasoning"；summary matched "evaluation"
   原始来源
3. ClawEnvKit: Automatic Environment Generation for Claw-Like Agents · Score 107
   title matched "agent"；summary matched "benchmark"；summary matched "evaluation"
   原始来源
4. MASS-RAG: Multi-Agent Synthesis Retrieval-Augmented Generation · Score 107
   title matched "agent"；summary matched "reasoning"；summary matched "benchmark"
   原始来源
5. OneVL: One-Step Latent Reasoning and Planning with Vision-Language Explanation · Score 106
   title matched "reasoning"；summary matched "agent"；summary matched "benchmark"
   原始来源

《CoopEval: Benchmarking Cooperation-Sustaining Mechanisms and LLM Agents in Social Dilemmas》〔评测 / 方法〕：It is increasingly important that LLM agents interact effectively and safely with other goal-pursuing agents, yet, re…

1. CoopEval: Benchmarking Cooperation-Sustaining Mechanisms and LLM Agents in Social Dilemmas · Score 130
   title matched "agent"；title matched "benchmark"；summary matched "reasoning"
   原始来源
2. IE as Cache: Information Extraction Enhanced Agentic Reasoning · Score 124
   title matched "agent"；title matched "reasoning"；summary matched "benchmark"
   原始来源
3. QuantCode-Bench: A Benchmark for Evaluating the Ability of Large Language Models to Generate Executable Algorithmic Trading Strategies · Score 123
   title matched "benchmark"；summary matched "agent"；summary matched "alignment"
   原始来源
4. From Reactive to Proactive: Assessing the Proactivity of Voice Agents via ProVoice-Bench · Score 122
   title matched "agent"；summary matched "reasoning"；summary matched "benchmark"
   原始来源
5. An Axiomatic Benchmark for Evaluation of Scientific Novelty Metrics · Score 109
   title matched "benchmark"；title matched "evaluation"；has PDF
   原始来源

《GeoAgentBench: A Dynamic Execution Benchmark for Tool-Augmented Agents in Spatial Analysis》〔评测 / 应用 / 方法〕：The integration of Large Language Models (LLMs) into Geographic Information Systems (GIS) marks a paradigm shift…

1. GeoAgentBench: A Dynamic Execution Benchmark for Tool-Augmented Agents in Spatial Analysis · Score 162
   title matched "agent"；title matched "benchmark"；summary matched "reasoning"
   原始来源
2. HINTBench: Horizon-agent Intrinsic Non-attack Trajectory Benchmark · Score 127
   title matched "agent"；title matched "benchmark"；summary matched "evaluation"
   原始来源
3. Character Beyond Speech: Leveraging Role-Playing Evaluation in Audio Large Language Models via Reinforcement Learning · Score 120
   title matched "evaluation"；summary matched "agent"；summary matched "reasoning"
   原始来源
4. LongCoT: Benchmarking Long-Horizon Chain-of-Thought Reasoning · Score 112
   title matched "reasoning"；title matched "benchmark"；has PDF
   原始来源
5. Correct Prediction, Wrong Steps? Consensus Reasoning Knowledge Graph for Robust Chain-of-Thought Synthesis · Score 108
   title matched "reasoning"；summary matched "benchmark"；summary matched "evaluation"
   原始来源

《Parallax: Why AI Agents That Think Must Never Act》〔评测 / 应用 / 方法〕：Autonomous AI agents are rapidly transitioning from experimental tools to operational infrastructure, with projections that 80% of enterprise application…

1. Parallax: Why AI Agents That Think Must Never Act · Score 107
   title matched "agent"；summary matched "reasoning"；summary matched "evaluation"
   原始来源
2. Drawing on Memory: Dual-Trace Encoding Improves Cross-Session Recall in LLM Agents · Score 107
   title matched "agent"；summary matched "reasoning"；summary matched "benchmark"
   原始来源
3. Round-Trip Translation Reveals What Frontier Multilingual Benchmarks Miss · Score 106
   title matched "benchmark"；summary matched "reasoning"；summary matched "evaluation"
   原始来源
4. Towards Long-horizon Agentic Multimodal Search · Score 106
   title matched "agent"；summary matched "reasoning"；summary matched "benchmark"
   原始来源
5. QuarkMedSearch: A Long-Horizon Deep Search Agent for Exploring Medical Intelligence · Score 105
   title matched "agent"；summary matched "benchmark"；summary matched "evaluation"
   原始来源

《UniToolCall: Unifying Tool-Use Representation, Data, and Evaluation for LLM Agents》〔评测 / 数据 / 应用 / 方法〕：Tool-use capability is a fundamental component of LLM agents, enabling them to interact with external systems throu…

1. UniToolCall: Unifying Tool-Use Representation, Data, and Evaluation for LLM Agents · Score 145
   title matched "agent"；title matched "evaluation"；summary matched "reasoning"
   原始来源
2. General365: Benchmarking General Reasoning in Large Language Models Across Diverse and Challenging Tasks · Score 130
   title matched "reasoning"；title matched "benchmark"；summary matched "evaluation"
   原始来源
3. Collaborative Multi-Agent Scripts Generation for Enhancing Imperfect-Information Reasoning in Murder Mystery Games · Score 129
   title matched "agent"；title matched "reasoning"；summary matched "benchmark"
   原始来源
4. FM-Agent: Scaling Formal Methods to Large Systems via LLM-Based Hoare-Style Reasoning · Score 127
   title matched "agent"；title matched "reasoning"；summary matched "evaluation"
   原始来源
5. From Translation to Superset: Benchmark-Driven Evolution of a Production AI Agent from Rust to Python · Score 126
   title matched "agent"；title matched "benchmark"；summary matched "evaluation"
   原始来源

收录 15 篇，重点包括《Paper Circle: An Open-source Multi-agent Research Discovery and Analysis Framework》、《Topological Characterization of Churn Flow and Unsupervised Correction to the Wu Flow-Regime Map in Small-Diameter Vertic…

1. Paper Circle: An Open-source Multi-agent Research Discovery and Analysis Framework
   原始来源
2. Topological Characterization of Churn Flow and Unsupervised Correction to the Wu Flow-Regime Map in Small-Diameter Vertical Pipes
   原始来源
3. The Character Error Vector: Decomposable errors for page-level OCR evaluation
   原始来源
4. MMEmb-R1: Reasoning-Enhanced Multimodal Embedding with Pair-Aware Selection and Adaptive Control
   原始来源
5. Toward Consistent World Models with Multi-Token Prediction and Latent Semantic Enhancement
   原始来源