Entities
open source project
ScienceWorld
- 2026-11-09researchPositiveIMP 70Researchers Introduce WM-PPO to Improve Agent Success with World Models
- 2026-11-05researchPositiveIMP 70AttnSR: A New Step-Level Reward Method for Long-Horizon Agentic Reinforcement Learning