qiqihezh / agentic-grpo-longhorizonView on GitHub
Fixing GRPO training collapse in long-horizon multi-tool agents. A lightweight PRM-Lite + LATA joint approach achieves +37% over vanilla GRPO on τ-bench airline (50-task, multi-turn).
90Jun 27, 2026Updated 2 weeks ago

Alternatives and similar repositories for agentic-grpo-longhorizon

Users that are interested in agentic-grpo-longhorizon are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.

Sorting:

Are these results useful?