richardyoung/ llama-3.1-8b-instruct-abliterated:Q4_K_M

756 1 month ago

An uncensored, abliterated build of Meta's **Llama 3.1 8B Instruct**. Abliteration surgically removes the model's learned refusal direction while preserving its general capabilities

... /
params
081256668599 · 111B
{
"num_ctx": 8192,
"stop": [
"<|start_header_id|>",
"<|end_header_id|>",
"<|eot_id|>"
]
}