Metraipolis › archive › dc_f5cc540b5ea3

Planning Commitment of LLM under Partial Observability — arXiv:2610.21108

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.21108
License (as archived): CC0-1.0
Archived: 2026-10-16T08:59:44.498Z
SHA-256 of the archived text: 2f965e6d2e9824f09d83ae3eca9b86a116eccc41f378b7c8e5b72edee2c954fb
Stable URL of this page: https://metraipolis.com/documento/dc_f5cc540b5ea3

How to cite

Planning Commitment of LLM under Partial Observability — arXiv:2610.21108 — arXiv (en), https://arxiv.org/abs/2610.21108 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f5cc540b5ea3 (Metraipolis archive dc_f5cc540b5ea3, sha256 2f965e6d2e9824f09d83ae3eca9b86a116eccc41f378b7c8e5b72edee2c954fb)

BibTeX

@misc{dcf5cc540b5ea3,
  title = {Planning Commitment of LLM under Partial Observability — arXiv:2610.21108},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.21108},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f5cc540b5ea3; sha256 2f965e6d2e9824f09d83ae3eca9b86a116eccc41f378b7c8e5b72edee2c954fb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21108. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM-based planning agents commit to plans of different lengths before interacting with a partially observable environment. Some aggressively commit to a long plan based on assumptions about unobserved states, while others conservatively plan only over what they have observed. Both behaviors emerge implicitly in pre-trained models, whether the LLM outputs a plan directly or generates a PDDL formulation for a classical solver. Aggressive commitment risks costly failures, while conservative commitment risks excessive LLM calls. This choice pits safety against efficiency, a trade-off that has been neither characterized systematically nor made controllable. We investigate aggressive and conservative LLM planners and formalizers on two partially observable domains, where priors are provided as either explicit statistics or common-sense knowledge. Across three LLMs, both strategies reach similar accuracy. Conservative methods achieve fewer failed attempts (safety) at the cost of more API calls (efficiency). Aggressive methods fail more often but replan efficiently only when priors are informative. Interestingly, analysis shows that LLMs treat the provided priors nearly as facts, which risks failure when priors are unreliable. We also explore hybrid strategies with tunable hyperparameters that adjust the balance between safety and efficiency.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.21108 — 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 2026-10-16T08:44:17.701Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).