FactSim: Fact-Checking for Opinion Summarization
Leandro Anghinoni, Jorge Sanchez · 2026
Read FactSim: Fact-Checking for Opinion Summarization by Leandro Anghinoni, Jorge Sanchez (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-08709.txt.
Description
Summaries of product reviews can sound fluent while inventing or dropping key facts, and standard metrics miss those errors. Leandro Anghinoni and Jorge Sanchez propose FactSim, a fully automated method for checking factual consistency in opinion summarization. Their arXiv paper targets generative AI evaluation where large language models changed the game.
The method extracts factual claims from summaries and original reviews, then scores coverage and consistency by comparing claim sets. It rewards similar claims even when negated, paraphrased, or expanded, and correlates highly with human judgment against leading metrics. Readers get a practical score for summary faithfulness that goes beyond surface overlap.
Catalog record
| CREATOR | Leandro Anghinoni, Jorge Sanchez |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-08709