Gordic: interpretability of looped transformers likely no harder than deep models
gordic_aleksa · x · 2026-11-22
AI researcher Aleksa Gordić argues that interpretability of looped transformers is presumably no harder than that of very deep models, and that 'one way or another we're heading in that direction' — a concise trend take from a practitioner.
More from Research
- Lean Community Verifies 17M Lines of Code, Confirms Claude-Assisted Riemann Hypothesis Proof — Polymarket · 2026-11-22
- Kotoba CEO: Audio Tokenization Is the Core Research Problem in Real-Time Speech AI — lmthang · 2026-11-22
- Optogenetics could have been invented in the 1970s — how common are biology's 'delayed inventions'? — NikoMcCarty · 2026-11-22
- NeurIPS 2026 Paper π-Fold Uses Physics-Informed RL to Fix Atom-Level Errors in Folding Models — lmthang · 2026-11-22
- Soon the interesting bit in math is minimizing Kolmogorov complexity of proofs — gordic_aleksa · 2026-11-22
- Russian philosophy journal tackles 'Do LLMs think?' with a computational answer — abenitezburraco · 2026-11-22