PrecogUI: Proactive GUI Agents via Pre-cognitive Simulation and Experience RetrievalResearchSep 30, 2026
When Upstream Messages Override Correct Answers: A Controlled Study of Multi-Agent LLM CollaborationResearchSep 30, 2026
Harness Evolution as Learning: Approximation, Generalization, and Optimization Limits of Self-Improving Personal AgentsResearchSep 30, 2026
STAR-GRPO: Canonical Anchoring and Reliability-First Advantages against Representation-Dependent Reward HackingResearchSep 30, 2026