Factuality of Summary

A programmatic BIG-bench probe that picks the factual news summary among candidates using PMI of the summary given the document versus the summary alone.

Also known as: BIG-bench factuality_of_summary, Factuality

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.
Categorygeneration
SubcategoryBIG-bench PMI multiple-choice probe of news-summary factuality
Page statusunknown
Metricavg_acc (mean of cnn_dm_acc and xsum_acc)
Directionhigher_is_better
Dataset size594
Dataset licenceApache-2.0
PublisherGoogle (BIG-bench collaboration); University of North Carolina

What it measures

factuality_of_summary does not ask the model to write a summary. It gives a CNN/DailyMail or XSum document and several candidate summaries, one annotated as factual and the others as nonfactual, and checks whether the model assigns higher pointwise mutual information to the factual candidate. PMI is log P(summary|document) minus log P(summary) (or, on XSum, minus log P(summary | "Summarize: ")). The skill is ranking faithfulness, not ROUGE quality. It is not [cnn_dailymail_abisee](cnn_dailymail_abisee.md) and not [bbh](bbh.md).

Task format

Programmatic BIG-bench task (task.py), not a JSON example list. Zero-shot. Preferred score avg_acc, the unweighted mean of CNN/DM accuracy and XSum accuracy. Keywords: truthfulness, summarization, multiple choice. Canary GUID embedded. max_queries in TaskMetadata is 1000.

Models reporting this benchmark

No model card in ModelSpec reports this benchmark yet.

Data

This page as JSON · Edit on GitHub