Tag: 问答系统 (2 articles)

A comprehensive guide to evaluating long-context Q&A systems covering metrics, dataset construction, and benchmark reviews across narrative and technical domains.

Long-context Q&A systems face challenges like information overload and multi-hop reasoning, and evaluation should focus on answer faithfulness and helpfulness to enhance user experience.