> the features tend to have greater semantical overlap? This is true. The features closer together now have much stronger semantic overlap. You can watch how

> the features tend to have greater semantical overlap? This is true. The features closer together now have much stronger semantic overlap. You can watch how the weights self-organize in a GPT here: https://toponets.github.io/webpage_assets/banner_video.mp4 https://toponets.github.io/webpage_assets/banner_video.mp4 We're already studying the effects of topographic structure on polysemanticity.