SuperSem

Repository for the superlatives dataset from the NAACL 2025 paper: Superlatives in Context: Modeling the Implicit Semantics of Superlatives

Paper Information

Authors: Valentina Pyatkin, Bonnie Webber, Ido Dagan, Reut Tsarfaty
Conference: North American Chapter of the Association for Computational Linguistics (NAACL) 2025
Year: 2025

Overview

This repository contains the SuperSem dataset, a collection of superlative expressions annotated with their implicit semantic properties in context. The dataset was introduced in our NAACL 2025 paper "Superlatives in Context: Modeling the Implicit Semantics of Superlatives."

Dataset

The SuperSem dataset includes: the full file, called supersem_full.tsv, the train split, called supersem_train.tsv, the test split, supersem_test.tsv, and the dev split, supersem_dev.tsv.

The dataset is provided in TSV format with the following columns:

Column	Description
ID	Unique identifier for each sentence
source	Source domain or corpus of the original text
context	The surrounding paragraph or discourse context
sentence	The full sentence containing the superlative expression
sup_idx_sent	Index position of the superlative within the sentence
sup_idx_context	Index position of the superlative within the context
title	Title of the article or document (if applicable)
sup_class	Classification of the superlative type (only for instances from Wikipedia, source of the classes is Silke et al. 2009)
superlative	The extracted superlative expression
turn_idx (only for dialogue)	Turn index in dialogue (only for dialogue-based examples)
target	The entity that is being described by the superlative
CS	Comparison Set - the group of entities being compared
Anchor	Reference point or entity for the comparison
Property	The property or dimension along which entities are compared
Orientation	Whether the superlative indicates maximum or minimum value
Implicit	Binary annotation indicating if there are implicit elements
Amount	Quantity information related to the superlative (if applicable)
Rank	Ranking position indicated by the superlative (if applicable)

Note: Fields marked with -, or empty fields, indicate that the specific instance is not a superlative expression (but syntactically tagged as one).

Citation

If you use this dataset in your research, please cite our paper:

@inproceedings{pyatkin2025superlatives,
  title     = {Superlatives in Context: Modeling the Implicit Semantics of Superlatives},
  author    = {Pyatkin, Valentina and Webber, Bonnie and Dagan, Ido and Tsarfaty, Reut},
  booktitle = {Proceedings of the 2025 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies},
  year      = {2025},
  publisher = {Association for Computational Linguistics}
}

License

The SuperSem dataset is released under the Creative Commons Attribution 4.0 International License (CC BY 4.0). You are free to:

Share — copy and redistribute the material in any medium or format Adapt — remix, transform, and build upon the material for any purpose, even commercially

Under the following terms:

Attribution — You must give appropriate credit and citation, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.