4 ms·
I'd love to see it. What I like about your approach is that it circumvents the data problem. You don't need a dataset with review histories and flashcard conte
by ran3000 1y ago
I'd love to see it.
What I like about your approach is that it circumvents the data problem. You don't need a dataset with review histories and flashcard content in order to train a model.
- jarrett-ye 1y agoAndy also tested this idea. You can read his notes here: GPT-4 can probably estimate whether two flashcards are functionally equivalent https://notes.andymatuschak.org/zJ7PMGzjcgBUoPjLUHBF9jn https://notes.andymatuschak.org/zJ7PMGzjcgBUoPjLUHBF9jn GPT-4 can probably estimate whether one prompt will spoil retrieval of another https://notes.andymatuschak.org/zK9Y15pCnRMLoxUahLCzdyc https://notes.andymatuschak.org/zK9Y15pCnRMLoxUahLCzdyc