Answers are grounded in this paper and cite their sources.
Pre-registration of the second-model replication of the STGF-MAS stage-2 planning study (TravelPlanner, NATURAL PLAN Trip and Meeting). Before any run on the full test sets, it fixes the cells (QQ: Qwen builds and executes; DQ⁺ / QD⁺: cross cells on TravelPlanner; a Qwen T = 0.7 cell on Trip), the hypotheses with their tests (paired McNemar, bootstrap CI, Holm correction), the decision rules for the paper and the scoring rules. The Russian protocol is authoritative; README_EN.md is an English summary.
No takes yet. Share an insight, caveat, or question.
Sergey Zubarev (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: