arXiv 论文提出无真值下的步骤级信用评估与有限执行回放方法

View PDF HTML (experimental)

Abstract:Step-level credit evaluation can conflate numerical association with ordering and branch recovery with position-selection value. We test these distinctions in ALFWorld with two independently seeded replay banks: 27,520 continuations over 172 fixed turns. At 16 continuations per arm per bank, cross-bank replay-contrast associations are Pearson 0.736 and pooled Spearman 0.330. On 30 historical failures, A-count-selected branches at confidence-/implicit-selected positions gain 8.54 percentage points over same-position factual reruns, but their 3.33-point advantage over frozen-random positions under the same alternative-selection rule remains unresolved. Numerical association alone does not establish reliable pooled ordering; branch recovery alone does not establish better score-informed position selection. Historical judge alignment is modestly positive with an unresolved archived-control increment. Same-cohort C/D information-source comparisons and a short, sparse-exposure training pilot provide scope checks; incremental hindsight training utility remains unestablished. Findings are conditional on the tested policies, intervention support and finite budgets.

Comments:	80 pages, 14 figures, including appendices. Substantially revised from v2; adds independent replay-bank evaluation, claim-matched recovery comparisons, and a training pilot
Subjects:	Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Cite as:	arXiv:2608.19760 [cs.LG]
	(or arXiv:2608.19760v3 [cs.LG] for this version)
	https://doi.org/10.48550/arXiv.2608.19760 arXiv-issued DOI via DataCite

Submission history

From: Haiyue Zhang [view email]
[v1] Thu, 20 Aug 2026 08:04:00 UTC (199 KB)
[v2] Tue, 25 Aug 2026 17:45:26 UTC (229 KB)
[v3] Sun, 25 Oct 2026 17:11:47 UTC (399 KB)