nbeerbower/QwQ-R1-abliterated-TIES-Qwen2.5-32B

Full fine-tune vs FuseAI/FuseO1-DeepSeekR1-QwQ-SkyT1-32B-Preview. Fine-tune: 98% of the weights changed, across 64 of 64 layers.

Tensors changed: lm_head 1/1, Embeddings 0/1, Norms 26/129, MLP 192/192, Attention 446/448

2 derivatives checked

• mradermacher/QwQ-R1-abliterated-TIES-Qwen2.5-32B-i1-GGUF: Quantization
• mradermacher/QwQ-R1-abliterated-TIES-Qwen2.5-32B-GGUF: Quantization

On the Hub · Downloads on Model Pulse