view post Post 11719 βοΈ AutoAbliterationI made a Colab notebook to automatically abliterate models.It's quite general, so you can do interesting stuff like blocking a given language in the model outputs.π» Colab: https://colab.research.google.com/drive/1RmLv-pCMBBsQGXQIM8yF-OdCNyoylUR1?usp=sharing See translation π 37 37 π₯ 16 16 β€οΈ 6 6 π 5 5 π 3 3 + Reply
view post Post 6216 βοΈ Gemma 3 AbliteratedI noticed that Gemma 3 was much more resilient to refusal removal than other models like Qwen 2.5.I experimented with different recipes and improved the abliteration technique I wrote about last year.It's still experimental but the refusal rate is super low in my tests. Enjoy! mlabonne/gemma-3-4b-it-abliterated mlabonne/gemma-3-12b-it-abliterated mlabonne/gemma-3-27b-it-abliterated See translation 1 reply Β· π 13 13 π 9 9 π 3 3 β€οΈ 2 2 + Reply
huihui-ai/DeepSeek-R1-Distill-Qwen-Coder-32B-Fusion-9010 Text Generation β’ Updated Feb 22 β’ 25 β’ 7
huihui-ai/Llama-3.3-70B-Instruct-abliterated Text Generation β’ Updated Dec 12, 2024 β’ 55k β’ β’ 40
netcat420/MFANN-Llama3.1-Abliterated-SLERP-TIES-V3 Text Generation β’ Updated Nov 26, 2024 β’ 8 β’ 2
huihui-ai/Qwen2.5-Coder-32B-Instruct-abliterated Text Generation β’ Updated Nov 25, 2024 β’ 507 β’ 32
Goekdeniz-Guelmez/Josiefied-Qwen2.5-14B-Instruct-abliterated-v4-gguf Text Generation β’ Updated Feb 23 β’ 448 β’ 12