用主题建模优化大模型，提升毒性文本识别准确率。
Leveraging Large Language Models and Topic Modeling for Toxicity Classification
- 结合主题建模微调BERTweet和HateBERT，针对性提升分类性能。
- 在特定话题上微调后，F1分数显著优于GPT-4等主流模型。
- 揭示了大模型在毒性识别中存在偏差，适合内容安全研究者参考。
内容审核与毒性分类具有重要社会意义。然而研究表明，主流分类模型存在放大或弱化偏见的问题，可能忽视或不利某些边缘群体。研究指出标注者立场会影响黄金标准标签，进而使模型学习到标注者的偏见。为此，本文在数据集上对BERTweet和HateBERT进行微调，并结合主题建模策略开展内容审核。结果表明，针对特定话题微调后的模型，在F1分数上显著优于GPT-4、PerspectiveAPI和RewireAPI等主流分类模型。这进一步揭示了当前先进大语言模型在准确检测和理解文本毒性方面仍存在明显局限性。代码已开源：https://github.com/aheldis/Toxicity-Classification.git。
原文摘要 · Abstract (English)
Content moderation and toxicity classification represent critical tasks with significant social implications. However, studies have shown that major classification models exhibit tendencies to magnify or reduce biases and potentially overlook or disadvantage certain marginalized groups within their classification processes. Researchers suggest that the positionality of annotators influences the gold standard labels in which the models learned from propagate annotators' bias. To further investigate the impact of annotator positionality, we delve into fine-tuning BERTweet and HateBERT on the dataset while using topic-modeling strategies for content moderation. The results indicate that fine-tuning the models on specific topics results in a notable improvement in the F1 score of the models when compared to the predictions generated by other prominent classification models such as GPT-4, PerspectiveAPI, and RewireAPI. These findings further reveal that the state-of-the-art large language models exhibit significant limitations in accurately detecting and interpreting text toxicity contrasted with earlier methodologies. Code is available at https://github.com/aheldis/Toxicity-Classification.git.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。