YanLabs/gemma-3-27b-it-abliterated-normpreserve-v1

YanLabs/gemma-3-27b-it-abliterated-normpreserve-v1 is a 27 billion parameter causal language model developed by YanLabs, based on Google's Gemma-3-27b-it. This model has undergone norm-preserving biprojected abliteration to remove refusal behaviors, making it suitable for mechanistic interpretability research. It is specifically designed to study LLM safety mechanisms and abliteration techniques, with the Q8_0 quantization level recommended for preserving original capabilities without refusals.