Fine-Grained Perspectives: Modeling Explanations with Annotator-Specific Rationales
2026
Read Fine-Grained Perspectives: Modeling Explanations with Annotator-Specific Rationales (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-2604-21667.txt.
Description
Olufunke Sarumi and colleagues model why annotators disagree, treating each annotator's written reason as a fine-grained signal of that person's view. Their system predicts labels and reasons jointly, using who each annotator is and their profile facts through a User Passport that shapes inner vectors. It offers two explainer designs, one prompt-based and one that links label-model vectors into a text maker. Adding reason modeling beats a simple annotator-aware model on guesses, with the bridge design giving steadier label match.
Catalog record
| YEAR | 2026 |
|---|---|
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 32602 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.21667