vec2vec Hits 100% Top-1 Accuracy and 0.96 Cosine Similarity on Hardest Embedding Pairs
maier_ak · x · 2026-09-10
vec2vec attains cosine similarity up to 0.96, 100% top-1 accuracy, and mean rank 1 on the hardest cross-backbone pairs. On out-of-distribution data — tweets and clinical notes — it still exceeds 0.73 cosine and far outperforms random guessing, underscoring both the method's effectiveness and the privacy risk of stolen embedding stores.
More from Research
- FLASH: GPU-native deformable sim learns cloth folding in minutes on one RTX 5090 — ChongZzZhang · 2026-10-26
- LEMON: zero-shot LLM misbehavior monitor flags hallucinations from attention maps — yoavartzi · 2026-10-26
- Behind the EMNLP keynote: LMs as simulated learners to generate child language acquisition hypotheses — najoungkim · 2026-10-26
- The paper behind the EMNLP 2026 keynote on bigram 'simulated learners' — najoungkim · 2026-10-26
- Laying out DeepLIFT as a standalone computation graph: from sequence to attributions directly — jmschreiber91 · 2026-10-26
- Speculative Diffusion Paper Maps the Speed-Quality Pareto Frontier via Two Tradeoffs — UCLA · 2026-10-26