4 ms·
Direct Preference Optimization: Your Language Model Is a Reward Model
- rahimnathwani 3y agoPreviously mentioned by a few commenters: https://news.ycombinator.com/item?id=36139342 https://news.ycombinator.com/item?id=36139342 https://news.ycombinator.com/item?id=36423421 https://news.ycombinator.com/item?id=36423421 https://news.ycombinator.com/item?id=38659012 https://news.ycombinator.com/item?id=38659012 Newly interesting because Andrew Ng raved about it in his latest newsletter: https://www.deeplearning.ai/the-batch/issue-231/ https://www.deeplearning.ai/the-batch/issue-231/
- amai 3y agosee also https://hn.algolia.com/?q=Direct+Preference+Optimization https://hn.algolia.com/?q=Direct+Preference+Optimization