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

3,775 2 months 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

2 models

llama-3.1-8b-instruct-abliterated:latest
dd2b5a660554 • 4.9GB • 128K context window •
Text input • 2 months ago
llama-3.1-8b-instruct-abliterated:Q4_K_M
dd2b5a660554 • 4.9GB • 128K context window •
Text input • 2 months ago