Implicit Hierarchical GRPO: Decoupling Tool Invocation from Execution for Tool-Integrated Mathematical Reasoning

Documenti analoghi

• AWPO: Enhancing Tool-Use of Large Language Models through Adaptive Integration of Reasoning Rewards
  di: Lin, Zihan, et al.
  Pubblicazione: (2025)
• ResT: Reshaping Token-Level Policy Gradients for Tool-Use Large Language Models
  di: Lin, Zihan, et al.
  Pubblicazione: (2025)
• ResRL: Boosting LLM Reasoning via Negative Sample Projection Residual Reinforcement Learning
  di: Lin, Zihan, et al.
  Pubblicazione: (2026)
• When Self-Belief Misleads: Active Label Acquisition for Reinforcement Learning with Verifiable Rewards
  di: Wang, Li, et al.
  Pubblicazione: (2026)
• Divide-Then-Aggregate: An Efficient Tool Learning Method via Parallel Tool Invocation
  di: Zhu, Dongsheng, et al.
  Pubblicazione: (2025)