ASTRA-QA

Scores whether a RAG system covers required topics and avoids curated unsupported claims on 869 abstract questions over papers and news.

Also known as: AbSTRAct Question Answering over documents

unassessed

This page is a discovery lead. Nobody has yet assessed it against the catalogue contract, so it carries no disposition. Absence of evidence here is not evidence of staleness.
Categoryknowledge
Subcategoryabstract document QA for retrieval-augmented generation
Page statusactive
Metrictopic F1 (T-F1); also topic precision/recall and hallucination rates
Directionhigher_is_better
Unit%
Dataset size869
Dataset licenceODC-By-1.0 for annotations and metadata; source documents keep their own licences
PublisherThe Chinese University of Hong Kong, Shenzhen (School of Data Science), with Data Science Group, Huolala

What it measures

ASTRA-QA (AbSTRAct Question Answering over documents) tests whether a retrieval-augmented system can synthesise a long answer from academic papers and news rather than extract a short fact. Questions cover five types: single-document summarisation, two-way comparison, multi-way comparison, thematic enumeration, and temporal reasoning. Each item is also run under three retrieval scopes (Simple, Middle, Hard) that grow the distractor pool while keeping the same topic-set reference. Scoring checks topic coverage against a curated answer set and matches against a curated hallucination set of plausible but unsupported topics.

Task format

Given a question and a retrieval corpus, the system returns a free-form abstractive answer. An LLM extractor (GPT-5.1 in the paper) turns that answer into a topic set, which is matched to the gold answer topics and the hallucination set.

Models reporting this benchmark

No model card in ModelSpec reports this benchmark yet.

Data

This page as JSON · Edit on GitHub