The twelve reported checkpoints span routed-expert pools of 32–512 and native selections of 4–10 routed experts per token, with model roles and parameter counts reported in Table 1.

The twelve reported checkpoints span routed-expert pools of 32–512 and native selections of 4–10 routed experts per token, with model roles and parameter counts reported in Table 1. · CiteArk

测量结果

已评估 0/60 项

获得支持 0有争议或冲突 0尚未评估 60

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
printedValue: 30.5B	30.5 billion parameters	—	尚未评估
printedValue: 3.3B	3.3 billion parameters	—	尚未评估
printedValue: 128	128 experts	—	尚未评估
printedValue: 8	8 experts	—	尚未评估
printedValue: 0	0 experts	—	尚未评估
printedValue: 80B	80 billion parameters	—	尚未评估
printedValue: 3B	3 billion parameters	—	尚未评估
printedValue: 512	512 experts	—	尚未评估
printedValue: 10	10 experts	—	尚未评估
printedValue: 1	1 experts	—	尚未评估
printedValue: 16.8B	16.8 billion parameters	—	尚未评估
printedValue: 2.75B	2.75 billion parameters	—	尚未评估

The twelve reported checkpoints span routed-expert pools of 32–512 and native selections of 4–10 routed experts per token, with model roles and parameter counts reported in Table 1.

报告指标与观测值

Assessment（0）

关联运行（0）

测量结果