Measurement
Gates that lie · part 2
Your Reference Fails Its Own Reference
A LoRA patch scored 89.06% on its worst prompt against a frozen 90% floor and looked FAILED. The unpatched base, run against the same floor for the first time, also failed at 89.06%, on a different prompt. The bar was unreachable by construction. Llama.cpp's own f16-KV configuration misses its own f32 reference the same way, at 5 of 7 tested lengths.
Read