# Lexical Features in Coreference Resolution: To be Used With Caution

arXiv: https://arxiv.org/abs/1704.06779 · HTML: https://mlsanity.com/paper/1704.06779/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 22 Apr 2017.

## Built on by (most cited first, 12 of 12)

- [State-of-the-art generalisation research in NLP: A taxonomy and review](https://mlsanity.com/paper/2210.03050/) (2022)
- [GENTLE: A Genre-Diverse Multilayer Challenge Set for English NLP and Linguistic Evaluation](https://mlsanity.com/paper/2306.01966/) (2023)
- [OntoGUM: Evaluating Contextualized SOTA Coreference Resolution on 12 More Genres](https://mlsanity.com/paper/2106.00933/) (2021)
- [Are Large Language Models Robust Coreference Resolvers?](https://mlsanity.com/paper/2305.14489/) (2023)
- [Major Entity Identification: A Generalizable Alternative to Coreference Resolution](https://mlsanity.com/paper/2406.14654/) (2024)
- [Automatic Fake News Detection: Are Models Learning to Reason?](https://mlsanity.com/paper/2105.07698/) (2021)
- [Challenges to Evaluating the Generalization of Coreference Resolution Models: A Measurement Modeling Perspective](https://mlsanity.com/paper/2303.09092/) (2023)
- [GUMsley: Evaluating Entity Salience in Summarization for 12 English Genres](https://mlsanity.com/paper/2401.17974/) (2024)
- [Anatomy of OntoGUM--Adapting GUM to the OntoNotes Scheme to Evaluate Robustness of SOTA Coreference Algorithms](https://mlsanity.com/paper/2110.05727/) (2021)
- [Separating Retention from Extraction in the Evaluation of End-to-end Relation Extraction](https://mlsanity.com/paper/2109.12008/) (2021)
- [Incorporating Singletons and Mention-based Features in Coreference Resolution via Multi-task Learning for Better Generalization](https://mlsanity.com/paper/2309.11582/) (2023)
- [Mah=an=ama: A Unique Testbed for Literary Entity Discovery and Linking](https://mlsanity.com/paper/2509.19844/) (2025)

## Abstract (arXiv)

Lexical features are a major source of information in state-of-the-art coreference resolvers. Lexical features implicitly model some of the linguistic phenomena at a fine granularity level. They are especially useful for representing the context of mentions. In this paper we investigate a drawback of using many lexical features in state-of-the-art coreference resolvers. We show that if coreference resolvers mainly rely on lexical features, they can hardly generalize to unseen domains. Furthermore, we show that the current coreference resolution evaluation is clearly flawed by only evaluating on a specific split of a specific dataset in which there is a notable overlap between the training, development and test sets.