Aletheia-Bench/BatchOnline-GRPO-7B

Partial fine-tune vs deepseek-ai/DeepSeek-R1-Distill-Qwen-7B. Partial change: 84% of the weights changed (attention, mlp, norms), in 28 of 28 layers.

Tensors changed: Embeddings 0/1, Attention 194/196, MLP 82/84, Norms 39/57, lm_head 0/1

2 derivatives checked

• mradermacher/BatchOnline-GRPO-7B-i1-GGUF: Quantization
• mradermacher/BatchOnline-GRPO-7B-GGUF: Same weights, converted

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse