LoRA Fine-tuning Efficiently Undoes Safety Training from Llama 2-Chat 70B — LessWrong
Produced as part of the SERI ML Alignment Theory Scholars Program - Summer 2023 Cohort, under the mentorship of Jeffrey Ladish. …
Jeffrey Ladish Seri Ml Alignment Theory Scholars Program Theory Scholars Program Ongoing Release While Llama Code Llama
Source: lesswrong.com