Gemma 3 0.27B 0.3B

User reports Gemma 3 270M at 321 tok/s on a Ryzen 7 7700 CPU, up from 132 tok/s, by replacing the output head with a two-level prefix head that sorts the vocabulary by byte order. Setup is a custom prefix head with 2,048 groups of 128 tokens, AVX2 only, one token at a time. The same grouping gives 27-33% lower KL to the full head than Google's Gemma 4 token ordering across Gemma 3 270M/1B, Gemma 4 E2B, Qwen3-0.6B and gpt-oss-20b. Caveats: one text domain, no downstream evals, no speculative-decoding acceptance numbers.