30K queries

Emerging

1papers using it

2026first seen

The '30K queries' dataset contains a set of queries used to evaluate the performance of dense retrieval models in e-commerce sponsored search by providing a structured training signal derived from the disagreements among heterogeneous retrieval systems.

🔎 Find this dataset

Papers using 30K queries (1)

Scaling Dense Retrieval with LLM-Annotated Training Data: Structured Mining and Progressive Curriculum for E-Commerce Sponsored Search2026