Verify with Caution: The Pitfalls of Relying on Imperfect Factuality Metrics

• URL: http://arxiv.org/abs/2501.14883v2
• Date: Thu, 30 Jan 2025 18:13:05 GMT
• Title: Verify with Caution: The Pitfalls of Relying on Imperfect Factuality Metrics
• Authors: Ameya Godbole, Robin Jia,
• Abstract summary: We re-evaluate five state-of-the-art factuality metrics on a collection of 11 datasets for summarization, retrieval-augmented generation, and question answering.<n>We find that these evaluators are inconsistent with each other and often misestimate system-level performance.
• Score: 22.84997018004618
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Improvements in large language models have led to increasing optimism that they can serve as reliable evaluators of natural language generation outputs. In this paper, we challenge this optimism by thoroughly re-evaluating five state-of-the-art factuality metrics on a collection of 11 datasets for summarization, retrieval-augmented generation, and question answering. We find that these evaluators are inconsistent with each other and often misestimate system-level performance, both of which can lead to a variety of pitfalls. We further show that these metrics exhibit biases against highly paraphrased outputs and outputs that draw upon faraway parts of the source documents. We urge users of these factuality metrics to proceed with caution and manually validate the reliability of these metrics in their domain of interest before proceeding.

Related papers

• Reranking-based Generation for Unbiased Perspective Summarization [10.71668103641552]
  We develop a test set for benchmarking metric reliability using human annotations.<n>We show that traditional metrics underperform compared to language model-based metrics, which prove to be strong evaluators.<n>Our findings aim to contribute to the reliable evaluation and development of perspective summarization methods.
  arXiv  Detail & Related papers  (2025-06-19T00:01:43Z)
• Position: All Current Generative Fidelity and Diversity Metrics are Flawed [58.815519650465774]
  We show that all current generative fidelity and diversity metrics are flawed.<n>Our aim is to convince the research community to spend more effort in developing metrics, instead of models.
  arXiv  Detail & Related papers  (2025-05-28T15:10:33Z)
• Goodhart's Law Applies to NLP's Explanation Benchmarks [57.26445915212884]
  We critically examine two sets of metrics: the ERASER metrics (comprehensiveness and sufficiency) and the EVAL-X metrics. We show that we can inflate a model's comprehensiveness and sufficiency scores dramatically without altering its predictions or explanations on in-distribution test inputs. Our results raise doubts about the ability of current metrics to guide explainability research, underscoring the need for a broader reassessment of what precisely these metrics are intended to capture.
  arXiv  Detail & Related papers  (2023-08-28T03:03:03Z)
• Evaluation of Faithfulness Using the Longest Supported Subsequence [52.27522262537075]
  We introduce a novel approach to evaluate faithfulness of machine-generated text by computing the longest noncontinuous of the claim that is supported by the context. Using a new human-annotated dataset, we finetune a model to generate Longest Supported Subsequence (LSS) Our proposed metric demonstrates an 18% enhancement over the prevailing state-of-the-art metric for faithfulness on our dataset.
  arXiv  Detail & Related papers  (2023-08-23T14:18:44Z)
• ROSCOE: A Suite of Metrics for Scoring Step-by-Step Reasoning [63.77667876176978]
  Large language models show improved downstream task interpretability when prompted to generate step-by-step reasoning to justify their final answers. These reasoning steps greatly improve model interpretability and verification, but objectively studying their correctness is difficult. We present ROS, a suite of interpretable, unsupervised automatic scores that improve and extend previous text generation evaluation metrics.
  arXiv  Detail & Related papers  (2022-12-15T15:52:39Z)
• ED-FAITH: Evaluating Dialogue Summarization on Faithfulness [35.73012379398233]
  We first present a systematic study of faithfulness metrics for dialogue summarization. We observe that most metrics correlate poorly with human judgements despite performing well on news datasets. We propose T0-Score -- a new metric for faithfulness evaluation.
  arXiv  Detail & Related papers  (2022-11-15T19:33:50Z)
• On the Limitations of Reference-Free Evaluations of Generated Text [64.81682222169113]
  We show that reference-free metrics are inherently biased and limited in their ability to evaluate generated text. We argue that they should not be used to measure progress on tasks like machine translation or summarization.
  arXiv  Detail & Related papers  (2022-10-22T22:12:06Z)
• On the Usefulness of Embeddings, Clusters and Strings for Text Generator Evaluation [86.19634542434711]
  Mauve measures an information-theoretic divergence between two probability distributions over strings. We show that Mauve was right for the wrong reasons, and that its newly proposed divergence is not necessary for its high performance. We conclude that -- by encoding syntactic- and coherence-level features of text, while ignoring surface-level features -- such cluster-based substitutes to string distributions may simply be better for evaluating state-of-the-art language generators.
  arXiv  Detail & Related papers  (2022-05-31T17:58:49Z)
• TRUE: Re-evaluating Factual Consistency Evaluation [29.888885917330327]
  We introduce TRUE: a comprehensive study of factual consistency metrics on a standardized collection of existing texts from diverse tasks. Our standardization enables an example-level meta-evaluation protocol that is more actionable and interpretable than previously reported correlations. Across diverse state-of-the-art metrics and 11 datasets we find that large-scale NLI and question generation-and-answering-based approaches achieve strong and complementary results.
  arXiv  Detail & Related papers  (2022-04-11T10:14:35Z)
• Investigating Crowdsourcing Protocols for Evaluating the Factual Consistency of Summaries [59.27273928454995]
  Current pre-trained models applied to summarization are prone to factual inconsistencies which misrepresent the source text or introduce extraneous information. We create a crowdsourcing evaluation framework for factual consistency using the rating-based Likert scale and ranking-based Best-Worst Scaling protocols. We find that ranking-based protocols offer a more reliable measure of summary quality across datasets, while the reliability of Likert ratings depends on the target dataset and the evaluation design.
  arXiv  Detail & Related papers  (2021-09-19T19:05:00Z)
• Understanding Factuality in Abstractive Summarization with FRANK: A Benchmark for Factuality Metrics [17.677637487977208]
  Modern summarization models generate highly fluent but often factually unreliable outputs. Due to the lack of common benchmarks, metrics attempting to measure the factuality of automatically generated summaries cannot be compared. We devise a typology of factual errors and use it to collect human annotations of generated summaries from state-of-the-art summarization systems.
  arXiv  Detail & Related papers  (2021-04-27T17:28:07Z)
• SummEval: Re-evaluating Summarization Evaluation [169.622515287256]
  We re-evaluate 14 automatic evaluation metrics in a comprehensive and consistent fashion. We benchmark 23 recent summarization models using the aforementioned automatic evaluation metrics. We assemble the largest collection of summaries generated by models trained on the CNN/DailyMail news dataset.
  arXiv  Detail & Related papers  (2020-07-24T16:25:19Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.