# Learning Sexism Detection Using Multi-Agent Perspectivist Preference Optimization

arXiv: https://arxiv.org/abs/2608.04056 · HTML: https://mlsanity.com/paper/2608.04056/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Social simulation and multi-agent debate, one of 16 directions in LLM agents and tool use.
- Builds on 9 of its 23 references that are in the map.
- First posted on arXiv 4 Aug 2026.

## Placed in

- LLM agents and tool use / [Social simulation and multi-agent debate](https://mlsanity.com/agents/social-simulation-and-multi-agent-debate/)
- LLM alignment / RLHF / preference optimisation / [Human values and AI feedback alignment](https://mlsanity.com/alignment/human-values-and-ai-feedback-alignment/)

## Builds on (most cited first, 9 of 9)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotations](https://mlsanity.com/paper/2110.05719/) (2021)
- [Direct Multi-Turn Preference Optimization for Language Agents](https://mlsanity.com/paper/2406.14868/) (2024)
- [Evaluating GRPO and DPO for Faithful Chain-of-Thought Reasoning in LLMs](https://mlsanity.com/paper/2512.22631/) (2025)
- [Assessing the Reliability of LLMs Annotations in the Context of Demographic Bias and Model Explanation](https://mlsanity.com/paper/2507.13138/) (2025)
- [Mars-PO: Multi-Agent Reasoning System Preference Optimization](https://mlsanity.com/paper/2411.19039/) (2024)

## Abstract (arXiv)

When people label text for sexism, they often disagree, and not because some of them are wrong: they genuinely perceive sexism differently. Most NLP systems discard this disagreement by collapsing it into a majority vote. We propose the Multi-Agent Perspectivist Preference Optimization (MAP-PO) framework to keep these different perspectives. On the EXIST 2024 dataset of labeled English and Spanish tweets, we first cluster annotators by their labeling behavior rather than their demographic attributes. We then fine-tune one Large Language Model agent per cluster to reproduce that cluster's annotation behavior, and coordinate the agents with preference optimization that combines individual and team-level rewards. We evaluate MAP-PO in four settings defined by two languages and two backbone language models, asking whether each agent reproduces the annotations of its own cluster and whether the agents together reproduce the majority label. Two findings hold in all four settings. First, without fine-tuning the agents behave almost identically, so cluster-specific training is necessary. Second, we show that training each agent only on the labels of its own cluster pushes the agents far beyond the clusters they should represent, while adding a shared team-level training signal consistently keeps each agent calibrated to its cluster.