Metraipolis › archive › dc_d8555767aeff

When Does Test-Time Search Help World Action Models? An Empirical Audit of Candidate Selection and Simulator Integrity — arXiv:2610.41970

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.41970
License (as archived): CC0-1.0
Archived: 2026-11-03T07:32:21.794Z
SHA-256 of the archived text: fa856172f631ef36875cb73d1a7ad043320abc07070291976aa457ceface666d
Stable URL of this page: https://metraipolis.com/documento/dc_d8555767aeff

How to cite

When Does Test-Time Search Help World Action Models? An Empirical Audit of Candidate Selection and Simulator Integrity — arXiv:2610.41970 — arXiv (en), https://arxiv.org/abs/2610.41970 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d8555767aeff (Metraipolis archive dc_d8555767aeff, sha256 fa856172f631ef36875cb73d1a7ad043320abc07070291976aa457ceface666d)

BibTeX

@misc{dcd8555767aeff,
  title = {When Does Test-Time Search Help World Action Models? An Empirical Audit of Candidate Selection and Simulator Integrity — arXiv:2610.41970},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41970},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d8555767aeff; sha256 fa856172f631ef36875cb73d1a7ad043320abc07070291976aa457ceface666d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41970. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

World-action models increasingly apply test-time scaling, spending extra inference computation on candidate selection and adaptive allocation to improve control. Such gains are credible only when candidates carry exploitable signal, nominal compute savings become real speedup, and evaluation stays reliable. We audit all three on Cosmos Policy and RoboCasa with a premise-first, pre-registered methodology. On clean data, a learned value selector is 1.66 pp worse than random; a hindsight oracle gains only 1.40 pp, with an episode-level interval spanning zero; and 82% of states show identical candidate outcomes, leaving little selection headroom. Token pruning cuts per-forward FLOPs by up to 47% but yields at most a measured 1.15$\times$ end-to-end speedup on the deployed path (1.25$\times$ on a decode-excluded path), short of a 1.5$\times$ target, and our adaptive-allocation experiment is inconclusive under a power gate. We further identify and fix a silent rendering corruption whose rate rises sharply along compute-heavy planning paths while leaving the official direct-policy result unaffected, and release a detector, fix, and scripts. These findings show why test-time scaling should be audited as a chain of premises (decision signal, realized efficiency, evaluationintegrity) rather than by success rate alone. Conclusions are scoped to the tested configuration.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.41970 — 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 2026-11-03T07:29:33.659Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).