https://huggingface.co/nicoboss/DeepSeek-R1-Distill-Llama-70B-Uncensored-v2
#590
by
nicoboss
- opened
This is a finetune of DeepSeek-R1-Distill-Llama-70B to make it uncensored. Unlike DeepSeek-R1-Distill-Llama-70B-Uncensored a much larger learning_rate was used making it heavily uncensored but still politically biased towards the Chinese narrative. In my next model the political bias will be removed as well.
I already put the source GGUF to /tmp/DeepSeek-R1-Distill-Llama-70B-Uncensored-v2.gguf on nico1.
In my next model the political bias will be removed as well.
nice
I already put the source GGUF
I moved it to /tmp/quant and will push the job to nico1.
mradermacher
changed discussion status to
closed