hjun1052

Wow, what a cool project! It struggles a bit with "쀍" tho... (but so does my phone’s Korean font lol)

If the model does autoregressive reasoning before the decision, doesn't that give up much of what a Jev-style model buys you (a single forward pass, cheap calibrated probabilities)? Or is the point mainly to keep the typed output and probability interface while getting better accuracy on harder cases?