Interpretability debate: does scaling make models denser, or just bigger?
JacobHHilton · x · 2026-10-18
A technical discussion on how interpretability scales. @banburismus distinguishes two conflated effects of increasing parameter count:
- More parameters → increasingly densely interconnected neurons/components
- More parameters → simply 'more stuff'
(1) would be a real obstacle for interpretability, since it makes interpreting individual units (e.g., single neurons) harder; (2) is a non-issue — just spend more tokens. JacobHHilton agrees with the distinction and responds in a thread.
More from Research
- Map of 171,250 RL papers shows deep RL and robotics dominate the field — CatAstro_Piyush · 2026-10-18
- KNOWS Benchmark: Best Web Agent Fully Succeeds on Fewer Than 3% of Complex Tasks — anmarasovic · 2026-10-18
- Open-source Engram memory engine, modeled on databases, matches Mem0 on BEAM — phecdaYm · 2026-10-18
- Mathematician: Pages of Routine Analysis Work Well, and AI Makes Writing Them Easier — jessi_cata · 2026-10-18
- Chinese researchers argue harness engineering is over: the eval set becomes the job — blaizedsouza · 2026-10-18
- Andrew Ng: 'AI for Science' will become meaningless as AI becomes every scientist's tool — andrewgwils · 2026-10-18