432-parameter model still unexplainable, sparking interpretability debate

Former OpenAI researcher Jacob Hilton showed a 432-parameter RNN that finds the second-largest number in a sequence yet still cannot be fully explained, while ARC released AlgZoo, tiny RNNs and transformers with 8–1,408 parameters. Hilton argued that peers' boundless optimism paints the public an unrealistic picture of interpretability's current state.

2026-10-16 ~ 2026-10-16 · 3 related posts

• ARC's AlgZoo: a 432-parameter model that interpretability still can't fully explain — JacobHHilton · 2026-10-16
• A 432-parameter RNN still defies full interpretability, ex-OpenAI researcher shows — JacobHHilton · 2026-10-16
• Interpretability drama: ARC researcher says a peer's optimism misleads the public — JacobHHilton · 2026-10-16