jeewan joga's picture

1 1 1

jeewan joga

jeewanjoga

·

AI & ML interests

None yet

Recent Activity

reacted to mlabonne's post with 👍 about 2 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

replied to mlabonne's post about 2 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

new activity 2 months ago

failspy/llama-3-70B-Instruct-abliterated:request: code for converting Qwen2.5-Coder-7B-Instruct models back to HF safetensors

View all activity

Organizations

None yet

jeewanjoga's activity

New activity in failspy/llama-3-70B-Instruct-abliterated 2 months ago

request: code for converting Qwen2.5-Coder-7B-Instruct models back to HF safetensors

#12 opened 2 months ago by