Metraipolis › archive › dc_c3fbd13ab74f

As Expected? Efficient LLM Planning via Expectation-Observation Consistency Verification — arXiv:2611.02438

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.02438
License (as archived)
CC0-1.0
Archived
2026-11-03T11:10:49.890Z
SHA-256 of the archived text
c3b1b51c0f9f462eee8e1e4842a9c2ca045044e57fe0dcbda07bb8a6e2625235
Stable URL of this page
https://metraipolis.com/documento/dc_c3fbd13ab74f

How to cite

As Expected? Efficient LLM Planning via Expectation-Observation Consistency Verification — arXiv:2611.02438 — arXiv (en), https://arxiv.org/abs/2611.02438 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c3fbd13ab74f (Metraipolis archive dc_c3fbd13ab74f, sha256 c3b1b51c0f9f462eee8e1e4842a9c2ca045044e57fe0dcbda07bb8a6e2625235)
BibTeX
@misc{dcc3fbd13ab74f,
  title = {As Expected? Efficient LLM Planning via Expectation-Observation Consistency Verification — arXiv:2611.02438},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.02438},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c3fbd13ab74f; sha256 c3b1b51c0f9f462eee8e1e4842a9c2ca045044e57fe0dcbda07bb8a6e2625235},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.02438. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM planning often relies on reactive executors (e.g., ReAct) that generate a single action at each step to remain grounded in the latest observations, incurring substantial inference cost. Batched executors reduce this cost by predicting multiple actions in a single model invocation, yet they are prone to compounding errors, as intermediate observations are ignored. We propose an expectation-observation consistency verification (EOCV) mechanism for batched executors, preserving grounding while improving efficiency. The mechanism generates expectations of action outcomes and compares them with the resulting observations. Consistency permits continued execution, whereas mismatches trigger early termination and replanning from the current observation. Experiments on text-based and multimodal benchmarks across multiple backbones show broadly comparable average task performance between EOCV and reactive executors, with a 29-35% reduction in average weighted token usage. Moreover, the mechanism is model-agnostic and remains effective when integrated into state-of-the-art planners.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2611.02438 — 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 2026-11-03T10:59:52.339Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).