一句话：提出低成本人在回路标注流程，用小型本地 LLM 加随机森林验证器完成 12.4 万条 GitHub 对话的毒性标注与实证分析。
开源讨论中的毒性言论会赶走贡献者、威胁项目可持续性，但以往 GitHub 毒性实证研究规模有限，结论的泛化性存疑；而毒性表达常隐式且依赖语境，纯人工标注与纯 LLM 标注都不可靠，难以扩展。本文提出一种人在回路（HITL）标注方法：单次调用小型本地 LLM，同时产出毒性预测与一组可解释的事件类别分数；再用轻量 Random Forest 验证器依据这些分数，挑出最可能被误标的少量对话，只在必要处引入人工复审。该验证器在效果上优于基于置信度和多 LLM 的基线，同时保持低标注成本。作者将流程应用于超过 124,000 条 GitHub issue 与 pull request 对话，用所得数据集检验此前小规模研究的关键结论，确认了其中一部分、限定了另一部分，并就不同开源项目中毒性的普遍程度、特征与动态给出新洞见。
Toxic interactions in open source discussions can alienate contributors and threaten project sustainability, yet prior empirical studies of GitHub toxicity have been limited in scale, raising questions about their generalizability. Scaling up is difficult because toxicity on GitHub is often implicit and context-dependent, making both fully manual annotation and LLM-based labeling unreliable. We present a human-in-the-loop (HITL) annotation methodology that makes large-scale, domain-calibrated toxicity labeling practical. A single call to a small, local LLM produces both a toxicity prediction and a set of interpretable event category scores. A lightweight Random Forest validator then uses those scores to flag the small subset of conversations most likely to be mislabeled, directing human review only where it is needed. The validator outperforms confidence-based and multi-LLM baselines while adding low annotation cost. We apply this pipeline to over 124,000 GitHub issue and pull request conversations. Using the resulting dataset, we evaluate key findings from prior small-scale research, confirming some and qualifying others, and present new insights into the prevalence, characteristics, and dynamics of toxicity across diverse open source projects.