On MATH-500, SPG and SPG+IM have similar step-to-step policy KL before step 2800, with IM marginally larger. From steps 2800–5000, when the baseline begins to collapse, the SPG baseline’s KL mean and volatility rise sharply, while SPG+IM remains lower and steadier.

On MATH-500, SPG and SPG+IM have similar step-to-step policy KL before step 2800, with IM marginally larger. From steps 2800–5000, when the baseline begins to collapse, the SPG baseline’s KL mean and volatility rise sharply, while SPG+IM remains lower and steadier. · CiteArk

测量结果

已评估 0/8 项

获得支持 0有争议或冲突 0尚未评估 8

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
run: SPG · period: before step 2800 · dataset: MATH-500	0.0052 KL divergence	—	尚未评估
run: SPG · period: before step 2800 · dataset: MATH-500	0.0078 KL divergence	—	尚未评估
run: SPG · period: steps 2800–5000 · dataset: MATH-500	0.0357 KL divergence	—	尚未评估
run: SPG · period: steps 2800–5000 · dataset: MATH-500	0.0365 KL divergence	—	尚未评估
run: SPG+IM · period: before step 2800 · dataset: MATH-500	0.0058 KL divergence	—	尚未评估
run: SPG+IM · period: before step 2800 · dataset: MATH-500	0.0089 KL divergence	—	尚未评估
run: SPG+IM · period: steps 2800–5000 · dataset: MATH-500	0.0221 KL divergence	—	尚未评估
run: SPG+IM · period: steps 2800–5000 · dataset: MATH-500	0.0292 KL divergence	—	尚未评估

On MATH-500, SPG and SPG+IM have similar step-to-step policy KL before step 2800, with IM marginally larger. From steps 2800–5000, when the baseline begins to collapse, the SPG baseline’s KL mean and volatility rise sharply, while SPG+IM remains lower and steadier.

报告指标与观测值

Assessment（0）

关联运行（0）

测量结果