SAE Research Resembles Older Papers
ninamiolane · x · 2026-07-15
The post revisits a discovery the author made two years ago: reading older papers by Bruno Olshausen's team, they found them strikingly similar to the superposition / SAE / interpretability research that has become popular in the Anthropic and MechInterp communities over the past couple of years.
Their initial reaction was: Are people missing citations?
But digging deeper, they realized this is actually a deeper historical lineage, not just a simple case of omitted references.
More from Research
- New prompt template aims to improve spatial reasoning and cut model laziness — legit_api · 2026-07-21
- Style-similarity analysis puts Kimi K3 closer to Claude Fable 5 than to K2.6 — soumitrashukla9 · 2026-07-21
- A GLP1R variant may explain stronger Ozempic weight loss, and the team built an agent workflow — julia_kiseleva · 2026-07-21
- Proceedings for the second geometry-grounded representation learning workshop are now online — erikjbekkers · 2026-07-21
- New survey maps how agentic systems are learning to improve themselves — SchmidhuberAI · 2026-07-21
- A curated TTS list for voice agents tracks latency, cancellation, and evals — mahimairaja · 2026-07-21