Decensored using a custom training script guided by activations, similar to ablation/"abliteration" scripts but not exactly the same approach.
I've found this effect to be stronger than most abliteration scripts, so please use responsibly etc etc.
The training script is released under the MIT license: https://github.com/nkpz/DeLMAT
- Downloads last month
- -
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support
Model tree for nkpz/Llama-3.1-8B-Instruct-Uncensored-DeLMAT
Base model
meta-llama/Llama-3.1-8B
Finetuned
meta-llama/Llama-3.1-8B-Instruct
Finetuned
unsloth/Meta-Llama-3.1-8B-Instruct