New preprint: cross-modal steering vectors transfer zero-shot both ways
benno_krojer · x · 2026-10-28
A new preprint on multimodal generalization shows cross-modal steering vectors transfer zero-shot in both directions—text generation to image generation and back—evidence for a shared conceptual space across modalities. Led by Donghao Zeng with Benno Krojer advising, co-authored by nsubramani and Tamar Rott Shaham.
More from Research
- K-Dense open-sources 177 science skills that turn coding agents into research scientists — alvelda · 2026-10-29
- Stanford team releases SOTA prompt optimization framework with unsupervised prompt performance measure — BrihiJ · 2026-10-29
- Debate set: Can we make AIs sentient — and should we? Solms vs. Humphrey — cccalum · 2026-10-29
- Samaya's 30B post-trained model beats frontier on financial research at 11x lower cost — maithra_raghu · 2026-10-29
- Scaling cross-situational word learning: 2,441 hours of child-perspective video from Stanford team — RichmanRonald · 2026-10-29
- Curiosity algorithms need world models plus locomotor primitives to replicate animal behavior — aran_nayebi · 2026-10-29