Episode
Self-Distilled Agentic Reinforcement Learning
- Podcast
- Daily Paper Cast
- Published
- May 16, 2026
- Duration seconds
- 1491
- Processing state
not_requested- Canonical source
- https://share.transistor.fm/s/896103fd
Actions
POST https://stenobird.com/v1/public/podcasts/daily-paper-cast-7079649/episodes/self-distilled-agentic-reinforcement-learning/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/daily-paper-cast-7079649/self-distilled-agentic-reinforcement-learning.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
🤗 Upvotes: 67 | cs.LG, cs.AI, cs.CL Authors: Zhengxi Lu, Zhiyuan Yao, Zhuowen Han, Zi-Han Wang, Jinyang Wu, Qi Gu, Xunliang Cai, Weiming Lu, Jun Xiao, Yueting Zhuang, Yongliang Shen Title: Self-Distilled Agentic Reinforcement Learning Arxiv: http://arxiv.org/abs/2605.15155v1 Abstract: Reinforcement learning (RL) has emerged as a central paradigm for post-training LLM agents, yet its trajectory-level reward signal provides only coarse supervision for long-horizon interaction. On-Policy Self-Distillation (OPSD) complements RL by introducing dense token-level guidance from a teacher branch augmented with privileged context. However, transferring OPSD to multi-turn agents proves problematic: compounding multi-turn instability destabilizes supervision, while skill-conditioned privileged guidance requires asymmetric treatment for negative teacher rejections may arise from imperfect skills retrieval or utilization. We introduce SDAR (Self-Distilled Agentic Reinforcement Learning), which treats OPSD as a gated auxiliary objective while keeping RL as the primary optimization backbone. SDAR maps detached token-level signals into a sigmoid gate, strengthening distillation on teacher-endorsed positive-gap tokens and softly attenuating negative teacher rejections. Across the Qwen2.5 and Qwen3 families on ALFWorld, WebShop, and Search-QA, SDAR substantially improves over GRPO (+9.4% on ALFWorld, +7.0% on Search-QA, +10.2% on WebShop-Acc), avoids the instability of naive GRPO+OPSD, and consistently outperforms hybrid RL--OPSD baselines across model scales.