Pith. sign in

REVIEW 2 cited by

Understanding Factual Errors in Summarization: Errors, Summarizers, Datasets, Error Detectors

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2205.12854 v2 pith:M4XBBGJS submitted 2022-05-25 cs.CL cs.AI

classification cs.CLcs.AI
keywords metricssummarizationerrorsfactualityerrormodelsfactualperformance
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

The propensity of abstractive summarization models to make factual errors has been studied extensively, including design of metrics to detect factual errors and annotation of errors in current systems' outputs. However, the ever-evolving nature of summarization systems, metrics, and annotated benchmarks makes factuality evaluation a moving target, and drawing clear comparisons among metrics has become increasingly difficult. In this work, we aggregate factuality error annotations from nine existing datasets and stratify them according to the underlying summarization model. We compare performance of state-of-the-art factuality metrics, including recent ChatGPT-based metrics, on this stratified benchmark and show that their performance varies significantly across different types of summarization models. Critically, our analysis shows that much of the recent improvement in the factuality detection space has been on summaries from older (pre-Transformer) models instead of more relevant recent summarization models. We further perform a finer-grained analysis per error-type and find similar performance variance across error types for different factuality metrics. Our results show that no one metric is superior in all settings or for all error types, and we provide recommendations for best practices given these insights.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. TiEBe: Tracking Language Model Recall of Notable Worldwide Events Through Time

    cs.CL 2025-01 conditional novelty 6.0 of 10

    A new benchmark, TiEBe, measures LLM recall of notable events across time, regions, and languages, and finds large geographic disparities correlated with GDP, HDI, and schooling.

  2. SummExecEdit: A Factual Consistency Benchmark in Summarization with Executable Edits

    cs.CL 2024-12 conditional novelty 6.0 of 10

    A new benchmark built with executable phrase-level edits shows that most LLMs detect and explain factual inconsistencies in summaries only weakly, with the best model scoring 0.49 on the joint task.

Pith tools