Metraipolis › archive › dc_171e69cf98d9

Self-Assessment Reinforcement Learning for Search-State Awareness in Retrieval-Augmented Large Language Models — arXiv:2610.30435

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.30435
License (as archived)
CC0-1.0
Archived
2026-10-23T15:14:54.291Z
SHA-256 of the archived text
14bf62f97b9ec9a566ff033ea0e955ff8d65e07e05e1702f5ea4ec166d4a8113
Stable URL of this page
https://metraipolis.com/documento/dc_171e69cf98d9

How to cite

Self-Assessment Reinforcement Learning for Search-State Awareness in Retrieval-Augmented Large Language Models — arXiv:2610.30435 — arXiv (en), https://arxiv.org/abs/2610.30435 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_171e69cf98d9 (Metraipolis archive dc_171e69cf98d9, sha256 14bf62f97b9ec9a566ff033ea0e955ff8d65e07e05e1702f5ea4ec166d4a8113)
BibTeX
@misc{dc171e69cf98d9,
  title = {Self-Assessment Reinforcement Learning for Search-State Awareness in Retrieval-Augmented Large Language Models — arXiv:2610.30435},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.30435},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_171e69cf98d9; sha256 14bf62f97b9ec9a566ff033ea0e955ff8d65e07e05e1702f5ea4ec166d4a8113},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.30435. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Recent advances in search agents have enabled large language models to tackle complex tasks by retrieving and synthesizing external information. However, existing search agents often lack search-state awareness, i.e., the ability to recognize what information has been obtained and what remains missing, leading to hallucinated reasoning and prematurely terminated search. To address this limitation, we propose Self-Assessment Reinforcement Learning (SARL), a reinforcement learning framework in which models learn to recognize their own search state during training and improve their search behavior accordingly, without requiring additional assessment at test time. SARL achieves this through a dual reward design that captures the corrective effect of each assessment while penalizing excessive reliance on it, jointly encouraging effective and efficient search trajectories. Under this setup, SARL achieves a 24.3% relative improvement over the strongest baseline, with average gains of 35.9% on multi-hop reasoning and 27.9% on deep research benchmarks against the strongest per-benchmark baselines. SARL further surpasses methods built on auxiliary supervision or multi-agent coordination, despite operating as a standalone RL framework with no inference-time scaffolding. Beyond final performance, SARL maintains training stability under extended search horizons, a regime where existing methods consistently degrade, enabling sustained gains as task complexity grows.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.30435 — 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 2026-10-23T15:09:08.518Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).