# Transfer Learning via Lexical Relatedness: A Sarcasm and Hate Speech Case Study

arXiv: https://arxiv.org/abs/2508.16555 · HTML: https://mlsanity.com/paper/2508.16555/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 27 references that are in the map.
- First posted on arXiv 22 Aug 2025.

## Builds on (most cited first, 8 of 8)

- [Latent Hatred: A Benchmark for Understanding Implicit Hate Speech](https://mlsanity.com/paper/2109.05322/) (2021)
- [Racial Bias in Hate Speech and Abusive Language Detection Datasets](https://mlsanity.com/paper/1905.12516/) (2019)
- [Understanding Abuse: A Typology of Abusive Language Detection Subtasks](https://mlsanity.com/paper/1705.09899/) (2017)
- [A BERT-Based Transfer Learning Approach for Hate Speech Detection in Online Social Media](https://mlsanity.com/paper/1910.12574/) (2019)
- [ETHOS: an Online Hate Speech Detection Dataset](https://mlsanity.com/paper/2006.08328/) (2020)
- [A Large Self-Annotated Corpus for Sarcasm](https://mlsanity.com/paper/1704.05579/) (2017)
- [Creating and Characterizing a Diverse Corpus of Sarcasm in Dialogue](https://mlsanity.com/paper/1709.05404/) (2017)
- [Decentralized Social Networks and the Future of Free Speech Online](https://mlsanity.com/paper/2406.06934/) (2024)

## Abstract (arXiv)

Detecting hate speech in non-direct forms, such as irony, sarcasm, and innuendos, remains a persistent challenge for social networks. Although sarcasm and hate speech are regarded as distinct expressions, our work explores whether integrating sarcasm as a pre-training step improves implicit hate speech detection and, by extension, explicit hate speech detection. Incorporating samples from ETHOS, Sarcasm on Reddit, and Implicit Hate Corpus, we devised two training strategies to compare the effectiveness of sarcasm pre-training on a CNN+LSTM and BERT+BiLSTM model. The first strategy is a single-step training approach, where a model trained only on sarcasm is then tested on hate speech. The second strategy uses sequential transfer learning to fine-tune models for sarcasm, implicit hate, and explicit hate. Our results show that sarcasm pre-training improved the BERT+BiLSTM's recall by 9.7%, AUC by 7.8%, and F1-score by 6% on ETHOS. On the Implicit Hate Corpus, precision increased by 7.8% when tested only on implicit samples. By incorporating sarcasm into the training process, we show that models can more effectively detect both implicit and explicit hate.