# THOS: A Benchmark Dataset for Targeted Hate and Offensive Speech

arXiv: https://arxiv.org/abs/2311.06446 · HTML: https://mlsanity.com/paper/2311.06446/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 6 of its 19 references that are in the map.
- First posted on arXiv 11 Nov 2023.

## Placed in

- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 6 of 6)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [BERTweet: A pre-trained language model for English Tweets](https://mlsanity.com/paper/2005.10200/) (2020)
- [Predicting the Type and Target of Offensive Posts in Social Media](https://mlsanity.com/paper/1902.09666/) (2019)
- [Multilingual and Multi-Aspect Hate Speech Analysis](https://mlsanity.com/paper/1908.11049/) (2019)
- [Latent Multi-task Architecture Learning](https://mlsanity.com/paper/1705.08142/) (2017)

## Built on by (most cited first, 4 of 4)

- [Guardians and Offenders: A Survey on Harmful Content Generation and Safety Mitigation of LLM](https://mlsanity.com/paper/2508.05775/) (2025)
- [Something Just Like TRuST : Toxicity Recognition of Span and Target](https://mlsanity.com/paper/2506.02326/) (2025)
- [AI Text Detectors and the Misclassification of Slightly Polished Arabic Text](https://mlsanity.com/paper/2511.16690/) (2025)
- [Towards Generalizable Generic Harmful Speech Datasets for Implicit Hate Speech Detection](https://mlsanity.com/paper/2506.16476/) (2025)

## Abstract (arXiv)

Detecting harmful content on social media, such as Twitter, is made difficult by the fact that the seemingly simple yes/no classification conceals a significant amount of complexity. Unfortunately, while several datasets have been collected for training classifiers in hate and offensive speech, there is a scarcity of datasets labeled with a finer granularity of target classes and specific targets. In this paper, we introduce THOS, a dataset of 8.3k tweets manually labeled with fine-grained annotations about the target of the message. We demonstrate that this dataset makes it feasible to train classifiers, based on Large Language Models, to perform classification at this level of granularity.