Beyond Precision: Importance-Aware Recall for Factuality Evaluation in Long-Form LLM Generation
Nazanin Jafari, James Allan, Mohit Iyyer · 2026
Read Beyond Precision: Importance-Aware Recall for Factuality Evaluation in Long-Form LLM Generation by Nazanin Jafari, James Allan, Mohit Iyyer (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-2604-03141.txt.
Description
Jafari, Allan, and Iyyer argue that checking long-form model answers for truth misses half the job when only precision is scored. Their framework also measures recall against reference facts built from outside sources, weighting each fact by relevance and salience. They find current models score far better on precision than recall, covering top facts well while leaving much relevant material out.
Catalog record
| CREATOR | Nazanin Jafari, James Allan, Mohit Iyyer |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 32249 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.03141