Metraipolis › archive › dc_358bd3bf9f1b
Learn Now, Use Next, Trust Later: Prequential Test-Time Learning for LLM Agents — arXiv:2609.35911
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.35911
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T16:24:58.397Z
- SHA-256 of the archived text
5f856aa90ced2d09569dc5ac354b6baa2d70138d358a94b2f8a38cc244ad6a5e- Stable URL of this page
https://metraipolis.com/documento/dc_358bd3bf9f1b
How to cite
Learn Now, Use Next, Trust Later: Prequential Test-Time Learning for LLM Agents — arXiv:2609.35911 — arXiv (en), https://arxiv.org/abs/2609.35911 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_358bd3bf9f1b (Metraipolis archive dc_358bd3bf9f1b, sha256 5f856aa90ced2d09569dc5ac354b6baa2d70138d358a94b2f8a38cc244ad6a5e)
BibTeX
@misc{dc358bd3bf9f1b,
title = {Learn Now, Use Next, Trust Later: Prequential Test-Time Learning for LLM Agents — arXiv:2609.35911},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.35911},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_358bd3bf9f1b; sha256 5f856aa90ced2d09569dc5ac354b6baa2d70138d358a94b2f8a38cc244ad6a5e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35911. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Adapting large language model agents during deployment requires not only retaining past experience, but also turning new observations into timely guidance. Many test-time learning methods, however, acquire knowledge from completed episodes. Feedback from an ongoing interaction may therefore not be distilled into knowledge soon enough to help the next decision. Acquiring knowledge at the granularity of individual transitions could reduce this delay, but raises a separate challenge: a rule that is useful within one episode may not be reliable enough to guide future episodes. Waiting for validation can forfeit immediate benefits, whereas unrestricted reuse can propagate accidental or misattributed guidance. We introduce StepLearn, a nonparametric framework that separates immediate use from persistent trust. It turns informative transitions into hypotheses that can guide the next step, while requiring prospective validation before reuse across episodes. Their predicted effects are checked against subsequent observations outside the source episodes, and only sufficiently supported hypotheses become available for persistent guidance. This process updates external knowledge while keeping all model parameters fixed. Over five rounds on WebArena-Lite and ALFWorld, StepLearn achieves average success rates of 59.9% and 84.0% with GPT-5-mini, and 57.8% and 88.1% with Qwen3.5-35B-A3B, respectively. It outperforms EvoTest, the strongest baseline, by 2.2-12.7 percentage points across the four settings. Learning dynamics further shows that these gains are not restricted to the final repetition, with advantages already present on first task attempts in most settings. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.AI). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.35911 — 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 2026-10-07T15:54:53.736Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).