AI RESEARCH
Robust Tool Use via Fission-GRPO: Learning to Recover from Execution Errors
arXiv CS.LG
•
ArXi:2601.15625v2 Announce Type: replace Large language models (LLMs) can call tools effectively, yet they remain brittle in multi-turn execution: after a tool-call error, smaller models often fall into repetitive invalid re-invocations instead of interpreting the feedback and recovering. This failure mode persists because current