BLaIR

Emerging

1papers using it

2024first seen

BLaIR is a comprehensive benchmark that evaluates large language models as semantic encoders in recommendation scenarios, containing a large-scale Amazon Reviews dataset with over 570 million reviews and 48 million items, and covering various recommendation tasks.

🔎 Find this dataset

Papers using BLaIR (1)

Bridging Language and Items for Retrieval and Recommendation: Benchmarking LLMs as Semantic Encoders2024 · 25 cites