Research · MarkTechPost ·
Stanford Researchers Introduce TRACE: A Capability-Targeted Agentic Training System That Turns Recurrent Agent Failures Into Synthetic RL Environment
Stanford researchers introduced TRACE, a system that identifies recurring capability gaps from agent trajectories and generates verifiable synthetic RL environments for targeted training. It uses LoRA adapters and token routing, reporting gains of 15.3 points on τ²-Bench and 73.2% Pass@1 on SWE-bench Verified.