FAMA boosts open-source LLM tool-use agents by up to 27% via failure trajectory analysis and orchestration of specialized agents for targeted context injection.
message”: “Message ID 3 deleted
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
FAMA: Failure-Aware Meta-Agentic Framework for Open-Source LLMs in Interactive Tool Use Environments
FAMA boosts open-source LLM tool-use agents by up to 27% via failure trajectory analysis and orchestration of specialized agents for targeted context injection.