Who is Harry Potter? Inside Microsoft Research’s Fine-Tuning Method for Unlearning Concepts in LLMs
pub.towardsai.net/who-is-harry-potter-inside-microsoft-researchs-fine-tuning-method-for-unlearning-concepts-in-llms-33dfe8e742a9
1 Users
0 Comments
4 Highlights
0 Notes
Tags
Top Highlights
This process identifies tokens whose probabilities have notably increased, indicating content-related tokens that should be avoided during generation
distinctive phrases from the target data are replaced with generic equivalents
Armed with these alternative labels, the model undergoes fine-tuning
What would a model unexposed to the Harry Potter books predict as the next token in this sentence
Ready to highlight and find good content?
Glasp is a social web highlighter that people can highlight and organize quotes and thoughts from the web, and access other like-minded people’s learning.