onchange to oninput and dropping the animation will make the slider feel much slicker. Our write up largely agrees with @Quintin Pope's summary, with the addition of training trajectory visualizations and an explanation of the MLP construction that solves modular addition.
A meta note that didn't make it into the article — with so many people looking into this problem over the last 18 months, I'm surprised this construction took so long to find. The modular addition task with a 1-layer MLP is about as simple as you can get![1]
Scaling mechanistic interpretability up to more complex tasks/models seems worth continuing to try, but I'm less sure extracting crisp explanations will be possible.[2] Even if we "solve" superposition, figuring the construction here — where there's no superposition in... (read 173 more words →)
Lots of custom d3 https://github.com/PAIR-code/ai-explorables/tree/master/source/grokking