The appendix reports every trained arm’s individual final validation loss and printed mean, including a preliminary TinyStories 10M experiment whose ordering matches the FineWeb 10M pattern: off-by-one and norm gate beat baseline, while the routing-slot control does not.

The appendix reports every trained arm’s individual final validation loss and printed mean, including a preliminary TinyStories 10M experiment whose ordering matches the FineWeb 10M pattern: off-by-one and norm gate beat baseline, while the routing-slot control does not. · CiteArk

测量结果

已评估 0/114 项

获得支持 0有争议或冲突 0尚未评估 114

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
arm: baseline · tier: TinyStories 10M · statistic: seed 0	1.5881 nats	—	尚未评估
arm: baseline · tier: TinyStories 10M · statistic: seed 1	1.587 nats	—	尚未评估
arm: baseline · tier: TinyStories 10M · statistic: seed 2	1.5869 nats	—	尚未评估
arm: baseline · tier: TinyStories 10M · statistic: mean	1.5873 nats	—	尚未评估
arm: off-by-one · tier: TinyStories 10M · statistic: seed 0	1.5852 nats	—	尚未评估
arm: off-by-one · tier: TinyStories 10M · statistic: seed 1	1.5826 nats	—	尚未评估
arm: off-by-one · tier: TinyStories 10M · statistic: seed 2	1.5835 nats	—	尚未评估
arm: off-by-one · tier: TinyStories 10M · statistic: mean	1.5838 nats	—	尚未评估
arm: renorm twin · tier: TinyStories 10M · statistic: seed 0	1.588 nats	—	尚未评估
arm: renorm twin · tier: TinyStories 10M · statistic: seed 1	1.5869 nats	—	尚未评估
arm: renorm twin · tier: TinyStories 10M · statistic: seed 2	1.5876 nats	—	尚未评估
arm: renorm twin · tier: TinyStories 10M · statistic: mean	1.5875 nats	—	尚未评估

The appendix reports every trained arm’s individual final validation loss and printed mean, including a preliminary TinyStories 10M experiment whose ordering matches the FineWeb 10M pattern: off-by-one and norm gate beat baseline, while the routing-slot control does not.

报告指标与观测值

Assessment（0）

关联运行（0）

测量结果