top | item 44608055 Gradient Descent on Token Input Embeddings 3 points| kp1197 | 7 months ago |lesswrong.com 1 comment order hn newest kp1197|7 months ago Does performing gradient descent on token input embeddings lead to interpretable results? And if not, why?
kp1197|7 months ago Does performing gradient descent on token input embeddings lead to interpretable results? And if not, why?
kp1197|7 months ago