ARC Researcher: We Understand Far Less Than 0.01% of Frontier Models' Cognition

ARC researcher Jacob Hilton argues that we understand far less than 0.01% of frontier models' internal cognition. He also revisited his 2020 Distill work on RL vision, noting its core attribution challenges remain unsolved.

2026-10-16 ~ 2026-10-16 · 2 related posts

• Revisiting Distill's 'Understanding RL Vision': attribution-based interpretability tricks that still matter — JacobHHilton · 2026-10-16
• ARC researcher: we understand far less than 0.01% of frontier LLMs' internal cognition — JacobHHilton · 2026-10-16