Hugging Face's logo
Marcus Hobbs's picture
TFLOPS icon 10.4 TFLOPS
2 40

Marcus Hobbs PRO

marcushobbs
21world's profile picture Mi6paulino's profile picture
·
https://wilsonic.co

AI & ML interests

SFT of domain expertise and conversation, edge inference, TTS, microtonality, software synthesis

Recent Activity

liked a model 2 days ago
OuteAI/OuteTTS-0.2-500M-GGUF
liked a dataset 7 days ago
allenai/WildChat-1M
liked a model 8 days ago
smpanaro/Llama-3.2-3B-Instruct-CoreML
View all activity

Organizations

MLX Community's profile picture

marcushobbs's activity

liked a model 2 days ago
liked a dataset 7 days ago
liked 2 models 8 days ago
liked a dataset 19 days ago
liked 2 models 23 days ago
liked 2 models 28 days ago
upvoted a paper about 1 month ago
liked a dataset about 1 month ago
upvoted an article about 1 month ago
Reacted to mlabonne's post with ❤️🔥 about 1 month ago
view post
Post
16478
✂️ Uncensor any LLM with abliteration

I wrote an article about abliteration and how NeuralDaredevil-8B was created. Beyond removing alignment, I believe it's an interesting technique with a lot of potential. It's basically fine-tuning without retraining.

In this article, we see how it works, implement it in Google Colab, and heal the abliterated model to recover the performance drop due to this technique. The final model is an uncensored and high-quality model with the highest MMLU score on the Open LLM Leaderboard (8B category).

https://huggingface.co/blog/mlabonne/abliteration
·
liked 3 datasets about 1 month ago
liked a model about 1 month ago
liked a dataset about 1 month ago
liked a Space about 1 month ago