Unverified50% confidenceFactExact time
Mamba的隐状态维度d(state expansion factor)通常为16到64,而模型维度D可能为数千
1
Sources
50%
Confidence
Long-term
Relevance
8/12/2026
First Seen
Sources
Related Claims
UnverifiedScale AI在2024年的分析显示,部分模型在MMLU上的分数膨胀可能高达5-10个百分点,因测试集已泄露到训练数据中57% similarUnverified斯坦福大学HAI研究所2024年度AI指数报告显示,顶级闭源模型在MMLU基准上的最高分与第五名之间的差距,已从2022年的约8个百分点收窄至不足2个百分点54% similarUnverifiedAnalysis speculates that the model is an MoE model with 2 to 4 trillion parameters, activating only approximately 150 billion parameters per token.54% similarUnverified对于 198B 总参数的稀疏 MoE 模型,实际激活参数量可能仅为 20-40B 量级54% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/735119API
curl https://kongchang.com/api/v1/knowledge/claims/735119MCP
get_claim(id=735119)