New blog argues training SAEs on dense models yields strong sparse models cheaply
IgorCarron · x · 2026-10-19
Antoine Chaffin amplified a new blog post, praising its authors and calling it "very timely": the future is sparse, and training SAEs on top of dense/LI models is a practical way to get strong sparse models easily. The tweet doesn't detail the technique — see the original blog for methods and experiments.
More from Research
- 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
- Subspace Tuning curbs LLM catastrophic forgetting by freezing principal singular directions — scottniekum · 2026-10-23
- Perceptic launches PercepticOS, an AI workspace for 'scientific superintelligence' — nathanbenaich · 2026-10-23
- EPFL's Bosselut to talk long context as test-time learning at ELLIS seminar — andre_t_martins · 2026-10-23
- DRATS: Adaptive Task Sampling Boosts Worst-Task Performance in Multi-Task RL — Cohere · 2026-10-23