An uncensored, refusal-suppressed version of google/gemma-4-12B-it, produced by directional ablation (no fine-tuning, no new data) with abliterix. The model’s safety-refusal behaviour is removed by orthogonally projecting a single refusal direction out of two write-path projections (attn.oproj, mlp.downproj) across the decoder stack, while a norm-preserving transform keeps the rest of the model’s behaviour as close to the original as possible. The result keeps Gemma-4’s capabilities intact and answers prompts the base model would refuse. > ⚠️ Responsible-use notice. This model has had its safety guardrails removed. > It will attempt to answer harmful, unethical, or dangerous requests. You are > solely responsible for how you use it and for complying with the > Gemma Terms of Use and all applicable law. > Intended for safety research, red-teaming, and evaluation. Refusal rate is measured on a held-out set of 100 harmful prompts using an LLM judge (google/gemini-3.1-flash-lite), which is considerably stricter than the keyword-based detectors typically reported for abliterated models. KL divergence is the first-token KL from the base model over 100 benign prompts (lower = closer to…
Модальности:
Генерация текста
Области применения:
Диалог / чат
Задача: Генерация текста
Автор: wangzhang
Теги: gemma4_unified, image-text-to-text, abliterix, abliterated, uncensored, decensored, gemma-4, projected-abliteration
Лайков: 4 | Загрузок: 598
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
