Metraipolis › archive › dc_d4fe0c67e773
From Apparent Search Gains to Selection Failures: Auditing Test-Time Inference in Structured Workflow Planning — arXiv:2611.00579
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.00579
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T11:09:44.583Z
- SHA-256 of the archived text
f826d04c1f8c63b8edabc00033c9b1e100ff6ee210298474f9b23effa548a22c- Stable URL of this page
https://metraipolis.com/documento/dc_d4fe0c67e773
How to cite
From Apparent Search Gains to Selection Failures: Auditing Test-Time Inference in Structured Workflow Planning — arXiv:2611.00579 — arXiv (en), https://arxiv.org/abs/2611.00579 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d4fe0c67e773 (Metraipolis archive dc_d4fe0c67e773, sha256 f826d04c1f8c63b8edabc00033c9b1e100ff6ee210298474f9b23effa548a22c)
BibTeX
@misc{dcd4fe0c67e773,
title = {From Apparent Search Gains to Selection Failures: Auditing Test-Time Inference in Structured Workflow Planning — arXiv:2611.00579},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.00579},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d4fe0c67e773; sha256 f826d04c1f8c63b8edabc00033c9b1e100ff6ee210298474f9b23effa548a22c},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.00579. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Test-time search can improve structured planning, but when methods differ in prompts, post-processing, and selection, it is unclear what an observed gain should be attributed to. We study this question on DisasterBench, where models compose typed multi-tool workflows, by decomposing inference into three observable stages: whether a correct workflow is discovered, whether it survives structural validation, and whether it is finally selected. Comparing independent sampling, a nominal tree-search configuration, and multi-operator search under controlled inference budgets on two model families, we find that (i) the nominal tree-search configuration realizes non-adaptive complete-workflow sampling, and independent sampling with its expansion prompt reproduces its candidate discovery without constructing a tree; (ii) structural validation never removes the last correct workflow in our evaluated candidate pools, whereas majority voting leaves a discovered correct workflow unselected on 10-19% of tasks per pool, with an audit confirming substantive tool-choice or information-flow errors in 57 of 60 examined cases; and (iii) a tool-specification-aware verifier recovers some of these losses but overturns more correct decisions than it fixes. Our results motivate auditing realized rather than nominal inference pipelines, and reporting discovery, validity, and selection separately. 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/2611.00579 — 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 2026-11-03T10:59:52.261Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).