Can LLMs Recognize Toxicity? A Structured Investigation Framework and Toxicity Metric (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing toxicity metrics rely on encoder models trained on specific toxicity datasets, which are susceptible to out-of-distribution (OOD) problems and depend on the dataset’s definition of toxicity. |
| Approach: | They propose a robust metric grounded on LLMs to flexibly measure toxicity according to the given definition by analysing toxicity factors and intrinsic toxic attributes. |
| Outcome: | The proposed metric improves on conventional metrics by 12 points in the F1 score and shows that upstream toxicity significantly influences downstream metrics, suggesting that LLMs are unsuitable for toxicity evaluations within unverified factors. |
Similar Papers
Realistic Evaluation of Toxicity in Large Language Models (2024.findings-acl)
Copied to clipboard
| Challenge: | a large amount of data exposes large language models to toxicity and bias . prompt engineering can be easily bypassed with minimal prompt engineering. |
| Approach: | They propose a dataset that uses manually crafted prompts to nullify protective layers of large language models. |
| Outcome: | The proposed dataset shows that prompts can nullify protective layers of large language models. |
Dialectal Toxicity Detection: Evaluating LLM-as-a-Judge Consistency Across Language Varieties (2025.findings-emnlp)
Copied to clipboard
| Challenge: | toxicity detection of modern LLMs is underexplored due to dialectal differences. |
| Approach: | They evaluate toxicity detection by using LLMs as evaluators across diverse dialects . they create a multi-dialect dataset using synthetic transformations and human-assisted translations based on human-aided translations. |
| Outcome: | The proposed model shows that LLMs are sensitive to dialectal shifts and low-resource multilingual variation. |
♪ Something Just Like TRuST ♪ *: Toxicity Recognition of Span and Target (2026.findings-acl)
Copied to clipboard
| Challenge: | Toxic language is pervasive online, and because LLMs are trained on web data, it generates such content. |
| Approach: | They propose a large-scale dataset that synthesizes toxicity definitions and an annotation scheme . they use a rigorous human annotation process to evaluate the diversity of the annotations . |
| Outcome: | The proposed model outperforms existing models on three tasks and is not reliable. |
Enhancing LLM-based Hatred and Toxicity Detection with Meta-Toxic Knowledge Graph (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing methods to address toxicity issues with large language models are inadequate . lack of domain-specific knowledge leads to false negatives and excessive sensitivity to toxic speech limits freedom of speech. |
| Approach: | They propose a method that leverages graph search on a meta-toxic knowledge graph to enhance hatred and toxicity detection. |
| Outcome: | The proposed method lowers false positive rate and improves toxicity detection performance in out-of-domain scenarios. |
Unveiling the Implicit Toxicity in Large Language Models (2023.emnlp-main)
Copied to clipboard
| Challenge: | Recent studies focus on probing toxic outputs that can be easily detected with existing toxicity classifiers, but LLMs can generate diverse implicit toxic output that are difficult to detect via simply zero-shot prompting. |
| Approach: | They propose a reinforcement learning based attacking method to induce the implicit toxic outputs in large language models by fine-tuning toxicity classifiers. |
| Outcome: | The proposed method generates implicit toxic outputs that are difficult to detect via zero-shot prompting on five widely-adopted toxicity classifiers. |
Challenges in Detoxifying Language Models (2021.findings-emnlp)
Copied to clipboard
Johannes Welbl, Amelia Glaese, Jonathan Uesato, Sumanth Dathathri, John Mellor, Lisa Anne Hendricks, Kirsty Anderson, Pushmeet Kohli, Ben Coppin, Po-Sen Huang
| Challenge: | Prior work often relies on automatic evaluation of LM toxicity. |
| Approach: | They evaluate toxicity mitigation strategies for automated and human evaluations . they find human raters disagree with high automatic toxicity scores after strong toxicity reduction interventions . |
| Outcome: | The proposed methods reduce LM toxicity but lower coverage for marginalized texts . human raters disagree with high toxicity scores after strong toxicity reduction interventions . |
Aligned Probing: Relating Toxic Behavior and Model Internals (2026.tacl-1)
Copied to clipboard
| Challenge: | Language models (LMs) may produce toxic text that contains hate speech, insults, or vulgarity, even when prompted with innocuous text. |
| Approach: | They propose an interpretability framework that aligns the behavior of language models based on their outputs and internal representations. |
| Outcome: | The proposed framework bridges behavioral and internal perspectives for toxicity for the first time. |
ROBBIE: Robust Bias Evaluation of Large Generative Language Models (2023.emnlp-main)
Copied to clipboard
David Esiobu, Xiaoqing Tan, Saghar Hosseini, Megan Ung, Yuchen Zhang, Jude Fernandes, Jane Dwivedi-Yu, Eleonora Presani, Adina Williams, Eric Smith
| Challenge: | generative large language models (LLMs) are becoming more performant and prevalent . we need tools to measure and improve their fairness, authors say . |
| Approach: | They propose to compare 6 different prompt-based bias and toxicity metrics across 12 demographic axes and 5 families of generative large language models. |
| Outcome: | The proposed model can be tested on more datasets to better characterize and mitigate biases . the study compared 6 prompt-based bias and toxicity metrics across 12 demographic axes and 5 families of generative large language models. |
A Survey of Toxicity Mitigation Strategies for Multilingual Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models can reproduce and amplify toxic content, including hate speech, harassment, and bias. |
| Approach: | They propose a comprehensive survey of the many detoxification methods tailored to multilingual LLMs. |
| Outcome: | The proposed methods are based on data filtering, style transfer, expert-based logit steering, retrieval augmentation, and human feedback. |
ModelCitizens: Representing Community Voices in Online Safety (2025.emnlp-main)
Copied to clipboard
Ashima Suvarna, Christina A Chance, Karolina Naranjo, Hamid Palangi, Sophie Hao, Thomas Hartvigsen, Saadia Gabriel
| Challenge: | Existing toxic language detection models are trained on annotations that collapse diverse perspectives into a single ground truth. |
| Approach: | They propose to augment social media posts with conversational scenarios to reflect the impact of conversational context on toxicity. |
| Outcome: | The proposed model outperforms existing models on social media with conversational scenarios. |