With 256 denoising steps at completion length 256 so one token is revealed per step when computing priorities, IM still exceeds its corresponding base framework in every reported MATH-500 and Countdown accuracy cell. Relative to default-schedule IM, the single-token schedule changes the twelve cells by an average of approximately +0.2 accuracy points.

With 256 denoising steps at completion length 256 so one token is revealed per step when computing priorities, IM still exceeds its corresponding base framework in every reported MATH-500 and Countdown accuracy cell. Relative to default-schedule IM, the single-token schedule changes the twelve cells by an average of approximately +0.2 accuracy points. · CiteArk

测量结果

已评估 0/37 项

获得支持 0有争议或冲突 0尚未评估 37

保留同一结论下的完整结果组。已有评估排在前面；单项判断不替代整组结论。展开行查看条件、来源与历次证据。

每行是一项测量，分别显示论文报告、最近可用实测和累计证据状态。
指标与条件	论文报告	最近可用实测	证据判断
model: SPG · dataset: MATH-500 · generationLengthTokens: 128	34.6%	—	尚未评估
model: SPG · dataset: MATH-500 · generationLengthTokens: 256	39.6%	—	尚未评估
model: SPG · dataset: MATH-500 · generationLengthTokens: 512	42.6%	—	尚未评估
model: SPG · dataset: Countdown · generationLengthTokens: 128	68.8%	—	尚未评估
model: SPG · dataset: Countdown · generationLengthTokens: 256	70.7%	—	尚未评估
model: SPG · dataset: Countdown · generationLengthTokens: 512	70.3%	—	尚未评估
model: SPG+IM · dataset: MATH-500 · generationLengthTokens: 128	34.6%	—	尚未评估
model: SPG+IM · dataset: MATH-500 · generationLengthTokens: 256	40.4%	—	尚未评估
model: SPG+IM · dataset: MATH-500 · generationLengthTokens: 512	45.2%	—	尚未评估
model: SPG+IM · dataset: Countdown · generationLengthTokens: 128	81.6%	—	尚未评估
model: SPG+IM · dataset: Countdown · generationLengthTokens: 256	81.6%	—	尚未评估
model: SPG+IM · dataset: Countdown · generationLengthTokens: 512	80.9%	—	尚未评估

报告指标与观测值

Assessment（0）

关联运行（0）

测量结果