Agents & Tool Use · 2025
ReTool: Reinforcement Learning for Strategic Tool Use in LLMs
An outcome-driven RL method that teaches a reasoning model when and how to invoke a code interpreter mid-reasoning, yielding large efficiency gains and emergent code self-correction.
Editorial record
Plain-language summary
An outcome-driven RL method that teaches a reasoning model when and how to invoke a code interpreter mid-reasoning, yielding large efficiency gains and emergent code self-correction.
Knowledge graph
Relationships
Descendants
Conceptual ancestorEvidence: Strongly supported
ReAct: Synergizing Reasoning and Acting in LMs
Builds on the agents lineage in the archive
freshness sweep 2026
Source record
Provenance
- Record ID
- P-602
- Record created
- 2026-07-13
- Last reviewed
- 2026-07-14
- Record version
- 2
- https://arxiv.org/abs/2504.11536
- arXiv:2504.11536
Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.