Hugging Face's logo
Woohyeuk Lee's picture

Woohyeuk Lee

kweel
failspy's profile picture
·
https://kweel.github.io/quartz

AI & ML interests

What defines "quality" training data How does each pipeline of data cleaning affect model performance and replicability of original data

Recent Activity

updated a dataset 2 days ago
kweel/LocalLLaMA-GitHub-READMEs
published a dataset 2 days ago
kweel/LocalLLaMA-GitHub-READMEs
replied to mlabonne's post 8 months ago
✂️ 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
View all activity

Organizations

None yet

models 0

None public yet