This is a heavily interactive web application, and JavaScript is required. Simple HTML interfaces are possible, but that is not what this is.
Post
Ted Underwood
tedunderwood.com
did:plc:565ebob5f6hw33hjdkxty6qj
A different approach to interpretability. Instead of probing neurons and circuits, *tune* the models to provide "accurate, quantitative descriptions of their own internal processes during certain kinds of decision-making," and then show that this ability generalizes. #MLSky 🤖
https://arxiv.org/abs/2505.17120
2025-06-12T14:31:57.247Z