2 ms·
> This is very different from Gemma 4, where the 26B-A4B model is much worse on several benchmarks (e.g. Codeforces, HLE) than 31B. Wouldn't you totally expect
by zozbot234 7mo ago
> This is very different from Gemma 4, where the 26B-A4B model is much worse on several benchmarks (e.g. Codeforces, HLE) than 31B.
Wouldn't you totally expect that, since 26A4B is lower on both total and active params? The more sensible comparison would pit Qwen 27B against Gemma 31B and Gemma 26A4B against Qwen 35A3B.
- Hugsun 7mo agoThey're comparing Qwen's moe vs dense (smaller difference) against Gemma's moe vs dense (bigger difference). Your proposed alternative misses the point.
- zozbot234 7mo agoGemma's dense is bigger than its moe's total parameters. You could totally expect the moe to do terribly by comparison.