We rewrote one weight matrix to teach a 0.8B model conversation classificationgiga.ai1 point·namanbnsl··0 commentsOpen articleSaveView on HN