# Distinguishing Right from Wrong in Debates: Attribution Analysis of Chinese Harmful Memes

arXiv: https://arxiv.org/abs/2605.24344 · HTML: https://mlsanity.com/paper/2605.24344/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 39 references that are in the map.
- First posted on arXiv 23 May 2026.

## Builds on (most cited first, 25 of 25)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks](https://mlsanity.com/paper/2312.14238/) (2023)
- [Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection](https://mlsanity.com/paper/2310.11511/) (2023)
- [A Survey on LLM-as-a-Judge](https://mlsanity.com/paper/2411.15594/) (2024)
- [Annotators with Attitudes: How Annotator Beliefs And Identities Bias Toxic Language Detection](https://mlsanity.com/paper/2111.07997/) (2021)
- [MOMENTA: A Multimodal Framework for Detecting Harmful Memes and Their Targets](https://mlsanity.com/paper/2109.05184/) (2021)
- [A Survey on Retrieval-Augmented Text Generation for Large Language Models](https://mlsanity.com/paper/2404.10981/) (2024)
- [ROUGE 2.0: Updated and Improved Measures for Evaluation of Summarization Tasks](https://mlsanity.com/paper/1803.01937/) (2018)
- [Out of the BLEU: how should we assess quality of the Code Generation models?](https://mlsanity.com/paper/2208.03133/) (2022)
- [Decoding the Underlying Meaning of Multimodal Hateful Memes](https://mlsanity.com/paper/2305.17678/) (2023)
- [What do you MEME? Generating Explanations for Visual Semantic Role Labelling in Memes](https://mlsanity.com/paper/2212.00715/) (2022)
- [Beneath the Surface: Unveiling Harmful Memes with Multimodal Reasoning Distilled from Large Language Models](https://mlsanity.com/paper/2312.05434/) (2023)
- [Towards Comprehensive Detection of Chinese Harmful Memes](https://mlsanity.com/paper/2410.02378/) (2024)
- [Benchmark dataset of memes with text transcriptions for automatic detection of multi-modal misogynistic content](https://mlsanity.com/paper/2106.08409/) (2021)
- [A Hybrid Approach to Information Retrieval and Answer Generation for Regulatory Texts](https://mlsanity.com/paper/2502.16767/) (2025)
- [Emotion-Aware Multimodal Fusion for Meme Emotion Detection](https://mlsanity.com/paper/2403.10279/) (2024)
- [HYRR: Hybrid Infused Reranking for Passage Retrieval](https://mlsanity.com/paper/2212.10528/) (2022)
- [MemeGraphs: Linking Memes to Knowledge Graphs](https://mlsanity.com/paper/2305.18391/) (2023)
- [InsightVision: A Comprehensive, Multi-Level Chinese-based Benchmark for Evaluating Implicit Visual Semantics in Large Vision Language Models](https://mlsanity.com/paper/2502.15812/) (2025)
- [From Recognition to Reasoning: Advancing Multimodal Harmful Meme Detection via Chain-of-Thought Alignment](https://mlsanity.com/paper/2506.18919/) (2025)
- [A Web-Scale Analysis of the Community Origins of Image Memes](https://mlsanity.com/paper/2204.05439/) (2022)
- [Multimodal Fake News Video Explanation: Dataset, Analysis and Evaluation](https://mlsanity.com/paper/2501.08514/) (2025)

## Abstract (arXiv)

Research on harmful meme detection has garnered significant attention, resulting in the development of numerous datasets and methods. However, progress in detecting Chinese harmful memes lags considerably, primarily due to two challenges: first, accurately assessing a meme's harmfulness depends heavily on understanding deep cultural context; second, many memes are semantically ambiguous, making harmfulness highly subjective. To address these issues, we focus on the interpretable detection of Chinese harmful memes by constructing the first Chinese harmful meme explanation dataset, Ex-ToxiCN-MM. This dataset offers opposing interpretations, categorized as "harmful" and "non-harmful", for each meme, aiming to rigorously evaluate a model's ability to discern and comprehend ambiguous, culturally grounded content. We built a specialized knowledge base of Chinese cultural concepts and offensive vocabulary to supply models with essential prior knowledge (C-HarmKB). To address the ambiguity and lack of background knowledge in meme attribution, we have developed a comprehensive attribution analysis framework, RIKE, which includes an Attribution Knowledge Enhancement module (AKE) and a Relative Intent Reasoning module (RIR). Extensive quantitative and qualitative experiments demonstrate that our method outperforms mainstream baseline models across multiple metrics in the task of attributing harmful memes in Chinese. The code, Ex-ToxiCN-MM dataset, and Chinese Harmful Semantic Knowledge Base (C-HarmKB) involved in this study have been open-sourced at https://github.com/wimiw123/Ex-ToxiCN-MM