# Mitigating Biases to Embrace Diversity: A Comprehensive Annotation Benchmark for Toxic Language

arXiv: https://arxiv.org/abs/2410.13313 · HTML: https://mlsanity.com/paper/2410.13313/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 29 of its 56 references that are in the map.
- First posted on arXiv 17 Oct 2024.

## Builds on (most cited first, 29 of 29)

- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://mlsanity.com/paper/2006.03654/) (2020)
- [Making Pre-trained Language Models Better Few-shot Learners](https://mlsanity.com/paper/2012.15723/) (2020)
- [Automated Hate Speech Detection and the Problem of Offensive Language](https://mlsanity.com/paper/1703.04009/) (2017)
- [TweetEval: Unified Benchmark and Comparative Evaluation for Tweet Classification](https://mlsanity.com/paper/2010.12421/) (2020)
- [Social Bias Frames: Reasoning about Social and Power Implications of Language](https://mlsanity.com/paper/1911.03891/) (2019)
- [True Few-Shot Learning with Language Models](https://mlsanity.com/paper/2105.11447/) (2021)
- [Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotations](https://mlsanity.com/paper/2110.05719/) (2021)
- [An Empirical Study of GPT-3 for Few-Shot Knowledge-Based VQA](https://mlsanity.com/paper/2109.05014/) (2021)
- [Predicting the Type and Target of Offensive Posts in Social Media](https://mlsanity.com/paper/1902.09666/) (2019)
- [Large Scale Crowdsourcing and Characterization of Twitter Abusive Behavior](https://mlsanity.com/paper/1802.00393/) (2018)
- [Can language models learn from explanations in context?](https://mlsanity.com/paper/2204.02329/) (2022)
- [Hate Speech Dataset from a White Supremacy Forum](https://mlsanity.com/paper/1809.04444/) (2018)
- [Are We Modeling the Task or the Annotator? An Investigation of Annotator Bias in Natural Language Understanding Datasets](https://mlsanity.com/paper/1908.07898/) (2019)
- [MultiVerS: Improving scientific claim verification with weak supervision and full-document context](https://mlsanity.com/paper/2112.01640/) (2021)
- [Assessing Social and Intersectional Biases in Contextualized Word Representations](https://mlsanity.com/paper/1911.01485/) (2019)
- [In Plain Sight: Media Bias Through the Lens of Factual Reporting](https://mlsanity.com/paper/1909.02670/) (2019)
- [Handling Bias in Toxic Speech Detection: A Survey](https://mlsanity.com/paper/2202.00126/) (2022)
- [Hate Lingo: A Target-based Linguistic Analysis of Hate Speech in Social Media](https://mlsanity.com/paper/1804.04257/) (2018)
- [Just Say No: Analyzing the Stance of Neural Dialogue Generation in Offensive Contexts](https://mlsanity.com/paper/2108.11830/) (2021)
- [Active Learning Principles for In-Context Learning with Large Language Models](https://mlsanity.com/paper/2305.14264/) (2023)
- [PERFECT: Prompt-free and Efficient Few-shot Learning with Language Models](https://mlsanity.com/paper/2204.01172/) (2022)
- [Not always about you: Prioritizing community needs when developing endangered language technology](https://mlsanity.com/paper/2204.05541/) (2022)
- [Adding Instructions during Pretraining: Effective Way of Controlling Toxicity in Language Models](https://mlsanity.com/paper/2302.07388/) (2023)
- [Impact of Training Dataset Size on Neural Answer Selection Models](https://mlsanity.com/paper/1901.10496/) (2019)
- [Contextual-Lexicon Approach for Abusive Language Detection](https://mlsanity.com/paper/2104.12265/) (2021)
- [The Effects of Data Size and Frequency Range on Distributional Semantic Models](https://mlsanity.com/paper/1609.08293/) (2016)
- [IndoToxic2024: A Demographically-Enriched Dataset of Hate Speech and Toxicity Types for Indonesian Language](https://mlsanity.com/paper/2406.19349/) (2024)

## Abstract (arXiv)

This study introduces a prescriptive annotation benchmark grounded in humanities research to ensure consistent, unbiased labeling of offensive language, particularly for casual and non-mainstream language uses. We contribute two newly annotated datasets that achieve higher inter-annotator agreement between human and language model (LLM) annotations compared to original datasets based on descriptive instructions. Our experiments show that LLMs can serve as effective alternatives when professional annotators are unavailable. Moreover, smaller models fine-tuned on multi-source LLM-annotated data outperform models trained on larger, single-source human-annotated datasets. These findings highlight the value of structured guidelines in reducing subjective variability, maintaining performance with limited data, and embracing language diversity. Content Warning: This article only analyzes offensive language for academic purposes. Discretion is advised.