SAGE

SAGE is a benchmark for scientific literature retrieval that contains 1,200 queries across four scientific domains and a retrieval corpus of 200,000 papers, used to evaluate the effectiveness of deep research agents and LLM-based retrievers in addressing complex queries.

Papers using SAGE (2)

Do Large Language Models Truly Understand Cross-cultural Differences?2025Are We on the Right Way to Assessing LLM-as-a-Judge?2025