Exploring Task Performance with Interpretable Models via Sparse Auto-Encoders

2 points | by PaulHoule 20 hours ago