Abliterated version of Qwen/Qwen3.5-122B-A10B with refusal direction removed. — Method: Refusal direction projection removal (Arditi et al., 2024) — Layers ablated: 5 (layers 43-47, covering both selfattn and linearattn/Mamba layers) — Tensors modified: 10 (oproj/outproj + qproj/inprojqkv per layer) — Alpha: 1.0 (full removal) — Measurement**: 64 harmful + 64 harmless prompts, strongest refusal signal at layer 48 (score: 78.4) — Type: Mixture of Experts (MoE) + Mamba hybrid attention — Total params: 122B — Active params: 10B per token (8/256 experts ro This model is intended for authorized security testing, CTF competitions, and educational purposes only.
Модальности:
Генерация текста
Области применения:
Диалог / чат
Задача: Генерация текста
Автор: Shifusen
Теги: qwen3_5_moe, quantization, fp8, abliterated, uncensored, pentest, qwen3.5, moe
Лайков: 3 | Загрузок: 223
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
