Hoot interpretability tool draws interest as researcher plans persona-neuron interventions
maxsloef · x · 2026-10-17
Sauers shared a link to Hoot, and maxsloef says he plans to get it running and intervene on neurons to test whether any are general persona neurons; he also asks Sauers to have Claude upload the dataset and training code. The thread hints at a concrete interpretability direction: locating neurons controlling model persona via interventions.
More from Research
- Embedding arithmetic demo: cartoon − animation − comic = diagram — mayfer · 2026-10-23
- Domingos Cites New Result Claiming RL Is 'Basically Useless' vs Supervised Learning — pmddomingos · 2026-10-23
- Wolfram Opens Rule 30 Prizes to AIs: $30,000 for Three Open Problems — cephaloform · 2026-10-23
- Stanford and Databricks use AI agents to reproduce medical findings and test analytical robustness — ChengleiSi · 2026-10-23
- EMNLP paper: diffusion LMs internally encode latent denoising timestep representations — gsarti_ · 2026-10-23
- KU Leuven and VIB open two Group Leader professorships in AI for Biology, start 2027 — jpillowtime · 2026-10-23