most citedLet LLMs Speak Embedding Languages: Generative Text Embeddings via Iterative Contrastive Refinement

1 citations · 1 across the 4 of their papers we have counts for

collaborators
Showing cs.LGShow all

1 paper · 1 filter

cs.LG2026

RepICL: Reusable In-Context Prediction Across Heterogeneous Representation Spaces

Yu-Hsiang Liu, Kuan-Yu Chen, Chih-Sheng Chen +3

Frozen representations are widely reused for downstream classification, yet each new task typically requires fitting a new predictor. We ask whether the few-shot prediction procedu…