Commonsense Knowledge with Negation: A Resource to Enhance Negation Understanding

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wang, Zijie, Rezaei, MohammadHossein, Rashid, Farzana, Blanco, Eduardo
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913053670375424
author Wang, Zijie
Rezaei, MohammadHossein
Rashid, Farzana
Blanco, Eduardo
author_facet Wang, Zijie
Rezaei, MohammadHossein
Rashid, Farzana
Blanco, Eduardo
contents Negation is a common and important semantic feature in natural language, yet Large Language Models (LLMs) struggle when negation is involved in natural language understanding tasks. Commonsense knowledge, on the other hand, despite being a well-studied topic, lacks investigations involving negation. In this work, we show that commonsense knowledge with negation is challenging for models to understand. We present a novel approach to automatically augment existing commonsense knowledge corpora with negation, yielding two new corpora containing over 2M triples with if-then relations. In addition, pre-training LLMs on our corpora benefits negation understanding.
format Preprint
id arxiv_https___arxiv_org_abs_2604_19921
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Commonsense Knowledge with Negation: A Resource to Enhance Negation Understanding
Wang, Zijie
Rezaei, MohammadHossein
Rashid, Farzana
Blanco, Eduardo
Computation and Language
Negation is a common and important semantic feature in natural language, yet Large Language Models (LLMs) struggle when negation is involved in natural language understanding tasks. Commonsense knowledge, on the other hand, despite being a well-studied topic, lacks investigations involving negation. In this work, we show that commonsense knowledge with negation is challenging for models to understand. We present a novel approach to automatically augment existing commonsense knowledge corpora with negation, yielding two new corpora containing over 2M triples with if-then relations. In addition, pre-training LLMs on our corpora benefits negation understanding.
title Commonsense Knowledge with Negation: A Resource to Enhance Negation Understanding
topic Computation and Language
url https://arxiv.org/abs/2604.19921