gpt-oss-120b config.json

Authoritative architecture fields for gpt-oss-120b: 36 layers, hidden 2880, 64 q heads, 8 kv heads, head_dim 64, 128 local experts, 4 per token, sliding_window 128, YaRN factor 32 from a 4096 base, max_position_embeddings 131072. | SOURCE AUDIT 2026-10-03: kind was 'spec-sheet' but the URL is a raw Hugging Face config.json served as text/plain, not a vendor spec sheet; set to 'repo'. The schema field it fills (layer/window/head counts) is machine-readable model metadata and should be read as such.