Metraipolis › archive › dc_e18903abfc19
Learning from Interaction Experience via Localized Context Distillation — arXiv:2610.21427
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.21427
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T09:03:21.436Z
- SHA-256 of the archived text
d469daae6b7c2e1569042b8d106408ea69b7a64f2edc4ba4d37ff8cd1cbbd894- Stable URL of this page
https://metraipolis.com/documento/dc_e18903abfc19
How to cite
Learning from Interaction Experience via Localized Context Distillation — arXiv:2610.21427 — arXiv (en), https://arxiv.org/abs/2610.21427 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e18903abfc19 (Metraipolis archive dc_e18903abfc19, sha256 d469daae6b7c2e1569042b8d106408ea69b7a64f2edc4ba4d37ff8cd1cbbd894)
BibTeX
@misc{dce18903abfc19,
title = {Learning from Interaction Experience via Localized Context Distillation — arXiv:2610.21427},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.21427},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e18903abfc19; sha256 d469daae6b7c2e1569042b8d106408ea69b7a64f2edc4ba4d37ff8cd1cbbd894},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21427. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Learning from interaction experience is essential for enabling large language models to continually improve during deployment. Context distillation offers a promising path: privileged information extracted from experience is provided to a teacher, whose behavior is then distilled into a student model that has no access to it. However, both existing forms of context distillation have limitations. Off-policy context distillation trains the student on teacher-generated trajectories and thus suffers from exposure bias; whereas on-policy context distillation trains on the student's own rollouts, whose states are often uninformative and fail to convey behaviors the student cannot yet produce. To resolve this dilemma, we introduce Localized Context Distillation (LoCD). LoCD first localizes the critical spans of a student trajectory that require revision. The teacher then regenerates only these spans from the student's own prefix, and the student is trained on the regenerated segments. LoCD thereby concentrates learning on the most informative states while keeping exposure bias low. Extensive experiments across math reasoning, code generation, and agent tasks demonstrate that LoCD consistently outperforms strong baselines and achieves more stable training with smaller and sparser parameter updates. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.21427 — 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 2026-10-16T08:44:17.743Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).