Metraipolis › archive › dc_aed558b8ff76

More Methods, Better Agents? A Reality Check on Self-Distillation in Agentic Reinforcement Learning — arXiv:2610.19789

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.19789
License (as archived): CC0-1.0
Archived: 2026-10-15T14:23:36.246Z
SHA-256 of the archived text: 3b7aa26d672d994e08e7cd3c24a08b6454dbeb3d8da2303ea1ba55addcb9dd50
Stable URL of this page: https://metraipolis.com/documento/dc_aed558b8ff76

How to cite

More Methods, Better Agents? A Reality Check on Self-Distillation in Agentic Reinforcement Learning — arXiv:2610.19789 — arXiv (en), https://arxiv.org/abs/2610.19789 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_aed558b8ff76 (Metraipolis archive dc_aed558b8ff76, sha256 3b7aa26d672d994e08e7cd3c24a08b6454dbeb3d8da2303ea1ba55addcb9dd50)

BibTeX

@misc{dcaed558b8ff76,
  title = {More Methods, Better Agents? A Reality Check on Self-Distillation in Agentic Reinforcement Learning — arXiv:2610.19789},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19789},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_aed558b8ff76; sha256 3b7aa26d672d994e08e7cd3c24a08b6454dbeb3d8da2303ea1ba55addcb9dd50},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19789. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Self-distillation in agentic reinforcement learning has produced a growing number of methods. However, existing evaluations often reuse published baseline scores under differing experimental conditions, provide few direct comparisons among contemporaneous methods, and give limited attention to training stability and computational cost. These limitations obscure the source of reported gains and the practical strengths and weaknesses of competing methods. Costly agent training further motivates a common basis for comparison. To address these gaps, we compare ten representative methods against GRPO on ALFWorld, ScienceWorld, Search-QA, and WebShop across three model sizes. Our study comprises 132 training runs totaling 12,124.4 recorded H800 GPU-hours, with repeated evaluations over eight rollout seeds for the pretrained backbones and all eleven trained methods. Beyond final task performance, we examine training progress, stability, and measured computational cost. Our evaluation yields three key findings. First, gains over GRPO remain limited in scope, with no method consistently outperforming it across settings. Second, strong final performance can mask unstable training. Third, additional training cost does not consistently translate into better performance. We establish this benchmark to support fair, comprehensive comparison and guide future research toward reliable and cost-effective improvements in agent training. Project page: this https URL

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.19789 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T13:53:35.905Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).