llama 3 在 arena 上的初步分数为 1199，基本上属于 gpt4 这一档

V2EX = way to explore

V2EX 是一个关于分享和探索的地方

For Existing Member Sign In

This topic created in 739 days ago, the information mentioned may be changed or developed.

参考 https://arena.lmsys.org/ 上的 Figure 3: Confidence Intervals on Model Strength (via Bootstrapping)

No Comments Yet