# Reliability-Aware Sexism Detection: Combining DPO with Annotator Agreement and Token-Level Confidence Scoring

arXiv: https://arxiv.org/abs/2608.12330 · HTML: https://mlsanity.com/paper/2608.12330/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in LLM preference optimization methods, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 15 of its 31 references that are in the map.
- First posted on arXiv 1 Jun 2026.

## Placed in

- LLM alignment / RLHF / preference optimisation / [LLM preference optimization methods](https://mlsanity.com/alignment/llm-preference-optimization-methods/)

## Builds on (most cited first, 15 of 15)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Selective Classification for Deep Neural Networks](https://mlsanity.com/paper/1705.08500/) (2017)
- [How Can We Know When Language Models Know? On the Calibration of Language Models for Question Answering](https://mlsanity.com/paper/2012.00955/) (2020)
- [Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotations](https://mlsanity.com/paper/2110.05719/) (2021)
- [KTO: Model Alignment as Prospect Theoretic Optimization](https://mlsanity.com/paper/2402.01306/) (2024)
- [Selective Question Answering under Domain Shift](https://mlsanity.com/paper/2006.09462/) (2020)
- [MultiVerS: Improving scientific claim verification with weak supervision and full-document context](https://mlsanity.com/paper/2112.01640/) (2021)
- [When the Majority is Wrong: Modeling Annotator Disagreement for Subjective Tasks](https://mlsanity.com/paper/2305.06626/) (2023)
- [Investigating Selective Prediction Approaches Across Several Tasks in IID, OOD, and Adversarial Settings](https://mlsanity.com/paper/2203.00211/) (2022)
- [TIS-DPO: Token-level Importance Sampling for Direct Preference Optimization With Estimated Weights](https://mlsanity.com/paper/2410.04350/) (2024)
- [Filtered Direct Preference Optimization](https://mlsanity.com/paper/2404.13846/) (2024)
- [Don't Waste a Single Annotation: Improving Single-Label Classifiers Through Soft Labels](https://mlsanity.com/paper/2311.05265/) (2023)
- [Re-examining Sexism and Misogyny Classification with Annotator Attitudes](https://mlsanity.com/paper/2410.03543/) (2024)
- [Exploring the Influence of Label Aggregation on Minority Voices: Implications for Dataset Bias and Model Training](https://mlsanity.com/paper/2412.04025/) (2024)

## Abstract (arXiv)

The detection of online sexism remains an open problem. Sexism detection is inherently subjective, yet most existing systems reduce multi-annotator labels to a single majority decision and treat all instances uniformly. This ignores two informative signals: annotator agreement and model uncertainty. We propose RA-DPO (Reliability-Aware Direct Preference Optimization), which integrates annotator agreement, model confidence, and a token-level uncertainty signal into a single reliability score. RA-DPO uses this score to select high-value preference pairs during training and to support inference-time abstention, which allows the model to trade coverage for accuracy. We evaluate RA-DPO on 6,920 multilingual posts from EXIST 2023, fine-tune OpenAI gpt-4o base via DPO, and validate on two open-weight 3B models (Llama, Qwen). Results show that training on the top 30% most reliable pairs matches full-data DPO, which indicates that reliability-aware selection can reduce training cost without sacrificing performance. At inference, selective prediction reaches 96.2% accuracy at 50% coverage in the true-agreement setting and 88.7% in the deployable predicted-agreement setting, both exceeding the 85.3% no-agreement baseline. These results suggest that accounting for annotation uncertainty is beneficial for both efficient training and reliable deployment in subjective classification.