Connor Leahy–EleutherAI, Conjecture
The Inside View
00:00
Interpretability
The reasony work o interpret ability. And we've just published our first technical post about interpretability, which is on how to defeat mind readers,. Well, yes. Ah, it's its a taxonomy and like, ride up of, like, all the ways we think nero networth could defeat interpretability tools. But yet, we also have some other cool work on polysementicity coming up, ofu nterpretability. So be see, the reason wer interest and interpretability is because we think it will be necessary, not because it's a solution to enlightenment. I don't think that the ti relses a solution to elitment. Bt
Transcript
Play full episode
