avichal comments on Learning to Interpret Weight Differences in Language Models