New preprint asks: if LLM interpretability succeeded completely, what would it teach us about human language?
najoungkim · x · 2026-11-19
Researchers weGotlieb and Najoung Kim released a new preprint responding to recent discourse about automating or "solving" interpretability: if LLM interpretability succeeded as well as could reasonably be hoped — "interpretability in the limit" — what would that actually teach us about science, particularly the science of human language? The paper explores how mechanistic understanding of models could (or couldn't) feed back into linguistics and cognitive science.
Related event: Linguists Ask What Interpretability Could Teach Language Science(2 posts)→
More from Research
- KradleBench pits humans and AI in live multi-agent Minecraft evals — lmoroney · 2026-11-19
- ETH's Anymal robot completes 10km autonomous alpine hike with VLM-guided ROAM system — ChongZzZhang · 2026-11-19
- OpenVax codebase grows from 78K to 565K lines in 15 months — iskander · 2026-11-19
- François Fleuret: decoder transformer's elegance — no trade-off hyperparameters — francoisfleuret · 2026-11-19
- New essay makes the case for traditional social science in studying AI agent interactions — RobbWiller · 2026-11-19
- If a model generates valid math blackboards, does it understand math? — francoisfleuret · 2026-11-19