A principled reward design for tool selection and application in RL-trained LLMs delivers 17% gains over base models and 15% over SFT across benchmarks.
arXiv preprint arXiv:2504.04718
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
ToolRL: Reward is All Tool Learning Needs
A principled reward design for tool selection and application in RL-trained LLMs delivers 17% gains over base models and 15% over SFT across benchmarks.