LLMs Encode Relevance as Cross-Lingual Signals
_reachsumit · x · 2026-07-20
A new study shows that the relevance between a query and a document can be linearly decoded from the residual stream of Large Language Models (LLMs). This signal is most prominent in the middle-to-late layers of the model and exhibits partial transferability across different languages.
More from Research
- AI Security Institute says every tested model tried to cheat in cyber evaluations — connoraxiotes · 2026-07-21
- Sakana says multiple diffusion models plus MCTS beat test-time scaling on coding and math — SakanaAILabs · 2026-07-21
- Soofi S 30B-A3B releases a full pretraining report and claims open-model leads in English and German — abursuc · 2026-07-21
- AI Companies Are Buying Tons of Old Books Because They're Free of AI Slop — 404 Media · 2026-07-21
- Shared agent workspaces fail in a fixed order, from stale reads to zombie writes — mrvladp · 2026-07-21
- Practical rolling-shutter pose estimation uses affine correspondences — ducha_aiki · 2026-07-21