dolphinlabs/ llama3.2-3b-uncensored:latest

674 5 days ago

Built with Llama. This model is Meta's Llama 3.2 3B Instruct with refusal behavior largely removed using abliteration, aka fining the "refusal direction" in the model's activations and projects it out of the weights. No fine-tuning was involved.

... /
params
50020e23ef83 ยท 126B
{
"stop": [
"<|start_header_id|>",
"<|end_header_id|>",
"<|eot_id|>"
],
"temperature": 0.6,
"top_p": 0.9
}