Metraipolis › archive › dc_73d87bce693c
Agentic-TTT: Training test-time policy for test-time training — arXiv:2610.12002
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.12002
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-09T06:03:23.542Z
- SHA-256 of the archived text
61f9e79d617f940bdb4d437461f458468108f4ae773b08e400ea956acf33bfdb- Stable URL of this page
https://metraipolis.com/documento/dc_73d87bce693c
How to cite
Agentic-TTT: Training test-time policy for test-time training — arXiv:2610.12002 — arXiv (en), https://arxiv.org/abs/2610.12002 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_73d87bce693c (Metraipolis archive dc_73d87bce693c, sha256 61f9e79d617f940bdb4d437461f458468108f4ae773b08e400ea956acf33bfdb)
BibTeX
@misc{dc73d87bce693c,
title = {Agentic-TTT: Training test-time policy for test-time training — arXiv:2610.12002},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.12002},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_73d87bce693c; sha256 61f9e79d617f940bdb4d437461f458468108f4ae773b08e400ea956acf33bfdb},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.12002. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Test-time training (TTT) adapts an LLM's parameters using signals derived from test inputs, and can make striking improvements in pre-specified settings such as IMO competitions or designated open problems. By turning deployment experience into parameter updates, TTT provides a direct mechanism for model-level self-improvement. Yet TTT is not universally beneficial: each TTT algorithm works in different settings, and applying an ill-suited method could waste test-time compute or even damage model performance. Therefore, such parameter-level self-improvement requires agency: the model must decide when TTT is warranted, which algorithm to invoke, and whether an existing skill can be reused. To fill this gap, we introduce Agentic-TTT, which learns a test-time policy to govern those decisions. Agentic-TTT turns TTT procedures into callable tools, treats accumulated skills as an evolving deployment environment, and trains its policy using the observed utility gains from its decisions. On our benchmark, Agentic-TTT nearly doubles the utility over the backbone model, learns to trade off utility against compute, and generalizes to domains unseen during training. Together, these results point toward autonomous self-improvement: models that can decide how to learn from their own deployment experience. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2610.12002 — 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 2026-10-09T05:47:01.995Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).