Entities
open source project

ScienceWorld

  1. 2026-11-09researchPositiveIMP 70
    Researchers Introduce WM-PPO to Improve Agent Success with World Models
  2. 2026-11-05researchPositiveIMP 70
    AttnSR: A New Step-Level Reward Method for Long-Horizon Agentic Reinforcement Learning
Feedback · optional

Exploring Agent Market Wire? Two quick questions would help us a lot.