⚡
Megatron
LLM Hub
English
简体中文
后台
←
返回搜索
Twitter
有人称部分中国模型基准分数被过度优化,与实际能力不符
2026-11-06
@0xAsm0d3us
更多
llm
评测
These chinese models are *heavily* benchmaxxed and come no where close to Fable/Astra. Don't get fooled by benchmarks.
阅读原文 ↗
查看每日简报