Evaluating the Factuality of Zero-shot Summarizers Across Varied Domains
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911771399290880 |
|---|---|
| author | Ramprasad, Sanjana Krishna, Kundan Lipton, Zachary C Wallace, Byron C |
| author_facet | Ramprasad, Sanjana Krishna, Kundan Lipton, Zachary C Wallace, Byron C |
| contents | Recent work has shown that large language models (LLMs) are capable of generating summaries zero-shot (i.e., without explicit supervision) that, under human assessment, are often comparable or even preferred to manually composed reference summaries. However, this prior work has focussed almost exclusively on evaluating news article summarization. How do zero-shot summarizers perform in other (potentially more specialized) domains? In this work we evaluate zero-shot generated summaries across specialized domains including biomedical articles, and legal bills (in addition to standard news benchmarks for reference). We focus especially on the factuality of outputs. We acquire annotations from domain experts to identify inconsistencies in summaries and systematically categorize these errors. We analyze whether the prevalence of a given domain in the pretraining corpus affects extractiveness and faithfulness of generated summaries of articles in this domain. We release all collected annotations to facilitate additional research toward measuring and realizing factually accurate summarization, beyond news articles. The dataset can be downloaded from https://github.com/sanjanaramprasad/zero_shot_faceval_domains |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_03509 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Evaluating the Factuality of Zero-shot Summarizers Across Varied Domains Ramprasad, Sanjana Krishna, Kundan Lipton, Zachary C Wallace, Byron C Computation and Language Artificial Intelligence Machine Learning Recent work has shown that large language models (LLMs) are capable of generating summaries zero-shot (i.e., without explicit supervision) that, under human assessment, are often comparable or even preferred to manually composed reference summaries. However, this prior work has focussed almost exclusively on evaluating news article summarization. How do zero-shot summarizers perform in other (potentially more specialized) domains? In this work we evaluate zero-shot generated summaries across specialized domains including biomedical articles, and legal bills (in addition to standard news benchmarks for reference). We focus especially on the factuality of outputs. We acquire annotations from domain experts to identify inconsistencies in summaries and systematically categorize these errors. We analyze whether the prevalence of a given domain in the pretraining corpus affects extractiveness and faithfulness of generated summaries of articles in this domain. We release all collected annotations to facilitate additional research toward measuring and realizing factually accurate summarization, beyond news articles. The dataset can be downloaded from https://github.com/sanjanaramprasad/zero_shot_faceval_domains |
| title | Evaluating the Factuality of Zero-shot Summarizers Across Varied Domains |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2402.03509 |