Metraipolis › archive › dc_579b146faa8f
Agents, Think Only When it Helps — arXiv:2611.05709
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.05709
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-05T18:16:44.482Z
- SHA-256 of the archived text
0f5f63c61806524e10f72bdeb8413f6f6b1d4bac659f48de6e010e7422d3048d- Stable URL of this page
https://metraipolis.com/documento/dc_579b146faa8f
How to cite
Agents, Think Only When it Helps — arXiv:2611.05709 — arXiv (en), https://arxiv.org/abs/2611.05709 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_579b146faa8f (Metraipolis archive dc_579b146faa8f, sha256 0f5f63c61806524e10f72bdeb8413f6f6b1d4bac659f48de6e010e7422d3048d)
BibTeX
@misc{dc579b146faa8f,
title = {Agents, Think Only When it Helps — arXiv:2611.05709},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.05709},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_579b146faa8f; sha256 0f5f63c61806524e10f72bdeb8413f6f6b1d4bac659f48de6e010e7422d3048d},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.05709. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
LLM agent distillation transfers agentic capabilities to student models by training on complete teacher trajectories that typically interleave thoughts and actions. We question how critical thought supervision is for agents' ultimate action performance. To answer this, we disentangle their impacts by independently perturbing teacher thoughts and actions during distillation. On widely adopted agent benchmarks across model scales, altering or discarding thoughts has little effect on student performance. We find, however, that thought supervision becomes important when information about the thought is not disclosed in the corresponding action during training. By identifying such discrepant cases, our findings challenge the common practice of distilling entire teacher trajectories for LLM agents, and offer a new insight that thought supervision can be a discernible choice for efficiency depending on whether the information of thought is disclosed in the action. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2611.05709 — 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 2026-11-05T18:12:48.997Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).