Original Model https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B

Heretic Ablitertion "[Trial 75] Refusals: 5/100, KL divergence: 0.00"

The Blashphemous Model has less refusals with a KL divergence of 0.01. I made this the heretic model to allow for making a LORA adapter, and to make merging better.

7B coming up next, If you release a GGUF please let me know so I can add it to my repo as well, I don't have the time to be learning how to do those conversions aat the moment.

Downloads last month
2
Safetensors
Model size
2B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support