OLMo-7B-Instruct-hf Parity F16 GGUF
Same-origin parity artifact derived from allenai/OLMo-7B-Instruct-hf.
This repo contains the high-fidelity f16 GGUF artifact used for mesh-llm backend parity validation against the corresponding MLX artifact.
Accepted local validation status:
- Exact prompts: GGUF and MLX align on semantic outputs, with shared family-level drift on strict one-word canaries
- Behavior smoke: 16 flagged prompts out of 80 on the MT-Bench-derived harness
Paired MLX repo:
meshllm/olmo-7b-instruct-hf-parity-bf16-mlx
