Jev model improves ICL demonstration retrieval via distillation into a lightweight MLP
kalyan_kpl · x · 2026-11-04
A paper applies the Jev model to demonstration selection for LLM in-context learning, assessing each candidate demonstration's utility for a given query. Running Jev at retrieval scale is expensive—every candidate needs an extra decision inference—so the authors distill its utility-estimation ability into a lightweight MLP that approximates Jev's probabilistic judgments, enabling efficient demonstration ranking with competitive retrieval quality.
More from Research
- Tencent's Hy4 rebuilds 7M-cell embryo in 1.5h, trains self-improving models — heyshrutimishra · 2026-11-04
- PixelBench: testing multimodal models pixel-by-pixel instead of with QA — bdsqlsz · 2026-11-04
- François Chollet: modeling brain wiring as an engineering problem constrained by genome size and time — CatAstro_Piyush · 2026-11-04
- OffsecCyber dataset combines HackerOne reports and CVEs to train pentesting LLMs — AccBalanced · 2026-11-04
- Monofact rate in training data governs LLM scientific discovery, preprint finds — CatAstro_Piyush · 2026-11-04
- qwenTPU compiles a TPU into Qwen3.5 weights, achieving exact integer matmul — CatAstro_Piyush · 2026-11-04