DevQuasar/xai-org.grok-2-GGUF

Same weights, converted (fewer layers) vs xai-org/grok-2. xai-org/grok-2 matches part of DevQuasar/xai-org.grok-2-GGUF (1 of its 963 tensors; the rest of DevQuasar/xai-org.grok-2-GGUF has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32), plus 962 new tensors.

Tensors changed: Embeddings 0/1

On the Hub · Downloads on Model Pulse