Skip to content
星际流动

Efficient Rationale-based Retrieval: On-policy Distillation from Generative Rerankers based on JEPA

发布
采集
学术前沿 7.0 分 — Revisits OPD formulation with theoretical analysis and empirical fixes for post-training
原文: arxiv.org

评分 7 · 来源: · 发布于 2026-04-28

评分依据:Revisits OPD formulation with theoretical analysis and empirical fixes for post-training