AI RESEARCH
TRACE: Capability-Targeted Agentic Training
arXiv CS.AI
•
ArXi:2604.05336v1 Announce Type: new Large Language Models (LLMs) deployed in agentic environments must exercise multiple capabilities across different task instances, where a capability is performing one or actions in a trajectory that are necessary for successfully solving a subset of tasks in the environment. Many existing approaches either rely on synthetic