# STEMTOX: From Collaborative Tags to Fine-Grained Toxic Meme Detection via Entropy-Guided Multi-Task Learning

arXiv: https://arxiv.org/abs/2508.04166 · HTML: https://mlsanity.com/paper/2508.04166/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 27 of its 45 references that are in the map.
- First posted on arXiv 6 Aug 2025.

## 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, 27 of 27)

- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks](https://mlsanity.com/paper/1610.02136/) (2016)
- [Automated Hate Speech Detection and the Problem of Offensive Language](https://mlsanity.com/paper/1703.04009/) (2017)
- [The Hateful Memes Challenge: Detecting Hate Speech in Multimodal Memes](https://mlsanity.com/paper/2005.04790/) (2020)
- [HateXplain: A Benchmark Dataset for Explainable Hate Speech Detection](https://mlsanity.com/paper/2012.10289/) (2020)
- [Recognize Anything: A Strong Image Tagging Model](https://mlsanity.com/paper/2306.03514/) (2023)
- [MultiVerS: Improving scientific claim verification with weak supervision and full-document context](https://mlsanity.com/paper/2112.01640/) (2021)
- [MOMENTA: A Multimodal Framework for Detecting Harmful Memes and Their Targets](https://mlsanity.com/paper/2109.05184/) (2021)
- [Tag2Text: Guiding Vision-Language Model via Image Tagging](https://mlsanity.com/paper/2303.05657/) (2023)
- [Detecting Harmful Memes and Their Targets](https://mlsanity.com/paper/2110.00413/) (2021)
- [Prompting for Multimodal Hateful Meme Classification](https://mlsanity.com/paper/2302.04156/) (2023)
- [Open-Set Image Tagging with Multi-Grained Text Supervision](https://mlsanity.com/paper/2310.15200/) (2023)
- [Probing LLMs for hate speech detection: strengths and vulnerabilities](https://mlsanity.com/paper/2310.12860/) (2023)
- [GOAT-Bench: Safety Insights to Large Multimodal Models through Meme-Based Social Abuse](https://mlsanity.com/paper/2401.01523/) (2024)
- [Pro-Cap: Leveraging a Frozen Vision-Language Model for Hateful Meme Detection](https://mlsanity.com/paper/2308.08088/) (2023)
- [Can't See the Forest for the Trees: Benchmarking Multimodal Safety Awareness for Multimodal LLMs](https://mlsanity.com/paper/2502.11184/) (2025)
- [Multimodal Language Models See Better When They Look Shallower](https://mlsanity.com/paper/2504.21447/) (2025)
- [BanglaAbuseMeme: A Dataset for Bengali Abusive Meme Classification](https://mlsanity.com/paper/2310.11748/) (2023)
- [Modularized Networks for Few-shot Hateful Meme Detection](https://mlsanity.com/paper/2402.11845/) (2024)
- [Vision-Language Models Create Cross-Modal Task Representations](https://mlsanity.com/paper/2410.22330/) (2024)
- [Exploring Large Language Models for Multi-Modal Out-of-Distribution Detection](https://mlsanity.com/paper/2310.08027/) (2023)
- [Exploring the Limits of Zero Shot Vision Language Models for Hate Meme Detection: The Vulnerabilities and their Interpretations](https://mlsanity.com/paper/2402.12198/) (2024)
- [Multi-modal Hate Speech Detection using Machine Learning](https://mlsanity.com/paper/2307.11519/) (2023)
- [Bridging Modalities: Enhancing Cross-Modality Hate Speech Detection with Few-Shot In-Context Learning](https://mlsanity.com/paper/2410.05600/) (2024)
- [PromptMTopic: Unsupervised Multimodal Topic Modeling of Memes using Large Language Models](https://mlsanity.com/paper/2312.06093/) (2023)
- [Toxicity Begets Toxicity: Unraveling Conversational Chains in Political Podcasts](https://mlsanity.com/paper/2501.12640/) (2025)
- [MMBERT: Scaled Mixture-of-Experts Multimodal BERT for Robust Chinese Hate Speech Detection under Cloaking Perturbations](https://mlsanity.com/paper/2508.00760/) (2025)

## Abstract (arXiv)

Memes, as a widely used mode of online communication, often serve as vehicles for spreading harmful content. However, limitations in data accessibility and the high costs of dataset curation hinder the development of robust meme moderation systems. To address this challenge, in this work, we introduce a first-of-its-kind dataset - TOXICTAGS consisting of 6,300 real-world meme-based posts annotated in two stages: (i) binary classification into toxic and normal, and (ii) fine-grained labelling of toxic memes as hateful, dangerous, or offensive. A key feature of this dataset is that it includes collaborative tags associated with the original posts, enhancing the context of each meme. In addition, we propose a novel entropy-guided multi-tasking framework -- STEMTOX -- that leverages these collaborative tags alongside visual and textual inputs within a robust classification framework. Experimental results show that incorporating these tags substantially enhances the performance of state-of-the-art VLMs in toxicity detection tasks. Our contributions offer a novel and scalable foundation for improved content moderation in multimodal online environments. Warning: Contains potentially toxic contents.