SEED: Self-Evolving Distillation for Agentic RL

A new paper introduces SEED, a self-evolving on-policy distillation framework designed for agentic reinforcement learning. It aims to overcome trajectory data limitations and significantly boost agent learning capabilities.

2026-07-17 ~ 2026-07-19 · 2 related posts

• SEED: Self-Evolving Agent Distillation — Jinyang Wu · 2026-07-17
• SEED: Self-Evolving RL for Agents — heghbalz · 2026-07-19