01. Harness Learning Enables Generalizable Test-Time Adaptation

85% worth reading · novelty 1.9/3 · new method · cs.CL

02. Failure-Transparent Agents: Benchmarking Post-Failure Reporting in Tool-Using Language Models

87% worth reading · novelty 1.6/3 · benchmark/eval · cs.AI

03. ReplayLens: Auditing Agents' Use of Outcomes

82% worth reading · novelty 2.0/3 · new method · cs.MA

04. Shockingly Simple Self-retrospection Improves Agentic Models Without RL

83% worth reading · novelty 1.9/3 · new method · cs.AI

05. Same Winners, Different Success Rates: Evaluating How LLM Agents Recover from Failures

81% worth reading · novelty 1.9/3 · benchmark/eval · cs.MA

06. MASTraceBench: Diagnosing Collaboration Gains through Proposal Trajectories in LLM-Based Multi-Agent Systems

81% worth reading · novelty 1.9/3 · benchmark/eval · cs.MA

07. TokenCast: Forecasting Token Consumption During LLM Agent Execution

83% worth reading · novelty 1.6/3 · new method · cs.AI

08. Reinforcing Agentic Creativity in Scientific Ideation with Night Science

79% worth reading · novelty 1.9/3 · new method · cs.AI

09. Prospective Interpretation Risk: Principled Communication Control Between LLMs

78% worth reading · novelty 2.0/3 · new method · cs.MA

10. FinAutoRubric: Expert-Guided Automatic Rubric Generation for Evaluating Financial Research Agents

81% worth reading · novelty 1.7/3 · benchmark/eval · cs.AI

11. Scaling Long-Form Story Generation via Narrative State Tracking

83% worth reading · novelty 1.3/3 · new method · cs.CL

12. Maat: Independent Deterministic Contract-Based Governance for Multi-Agent LLM Workflows

77% worth reading · novelty 1.7/3 · new method · cs.MA

13. Self-Adapting Group of Experts for Multi-Agent Reasoning

77% worth reading · novelty 1.6/3 · new method · cs.MA

14. From Migration to Calibration: Preserving Agent Capabilities across Models, Jurisdictions, and Scale

78% worth reading · novelty 1.6/3 · survey/position · cs.MA

15. Towards Communication-Efficient Social Intelligence in Language Agents

76% worth reading · novelty 1.6/3 · new method · cs.CL

16. Population Physics, Population Problems: Safety and Emergence in LLM Societies

72% worth reading · novelty 1.9/3 · new method · cs.MA

17. CEO Arena: Evaluating Long-Horizon Multi-Agent Decision-Making in Competitive Markets

73% worth reading · novelty 1.8/3 · benchmark/eval · cs.MA

18. QuanReview: Offline, Auditable Reconciliation of Human and LLM Span Annotations

81% worth reading · novelty 1.2/3 · applied system · cs.CL

19. KV-streams for Efficient Compaction in Agentic Reinforcement Learning

74% worth reading · novelty 1.7/3 · new method · cs.AI

20. Learning Native Reflection in Unified Models with Interleaved Reinforcement Learning

67% worth reading · novelty 1.9/3 · new method · cs.AI