Subliminal Steering presented at COLM: a more consistent take on subliminal learning
johnhewtt · x · 2026-10-06
At COLM, the author's lab is presenting Subliminal Steering, shown by George Morgulis (11:00-1:00, Franciscan A #78). The claim: if you study subliminal learning in language models, this steering-induced variant works far more consistently than existing approaches.
More from Research
- AI for Math Fund adds $17.1M from XTX Markets, backing 22 projects across 30 organizations — AlexKontorovich · 2026-10-06
- 62-person blind test of 48 LLM jokes shows models are getting funnier, Astra tops with 50%+ laughs — paraschopra · 2026-10-06
- Bayesian teaching dramatically improves probabilistic reasoning in LLMs, Nature Comms paper finds — tallinzen · 2026-10-06
- LiFT loops a DiT at inference: beats DiT-XL/2 with 52% less inference compute — cgmsnoek · 2026-10-06
- Researcher laments how slow humans look as algorithm design breakthroughs pile up — michaelchchoi · 2026-10-06
- Meta paper: dual coding agents cross-review lift correct patches from 45.8% to 62.5% — rohanpaul_ai · 2026-10-06