建模句子间关系，提升混合文档中AI生成文本的检测精度
SenFlow: Inter-Sentence Flow Modeling for AI-Generated Text Detection in Hybrid Documents
- 构建句子级图结构，融合上下文依赖关系进行联合预测
- 在MOSAIC数据集上实现跨领域迁移4.15%的宏F1提升
- 适合关注生成文本检测与对抗攻击的研究者
针对人类与大模型共同撰写的混合文档，现有句子级AI生成文本检测方法忽视句子间的依赖关系，且基准测试未涵盖最新生成器。我们构建了MOSAIC基准，包含16,000个来自PubMed和XSum的混合文档，由DeepSeek-V3.2和Kimi K2生成，并引入困惑度一致性过滤器以确保质量。将S-AGTD重新建模为文档级序列结构预测任务，提出SenFlow模型，通过图传播与线性链CRF解码，在单次文档遍历中捕捉句子间流动信息。SenFlow在MOSAIC上达到当前最优性能，跨领域迁移任务平均宏F1提升4.15个百分点，是三类难度递增协议中最难的一类。进一步发现，即使经困惑度过滤后显性线索被消除，不同生成器仍保留可被检测的句长差异。
原文摘要 · Abstract (English)
Sentence-level AI-generated text detection (S-AGTD) for hybrid documents, where humans and LLMs co-author one text, faces two gaps: existing methods classify each sentence in isolation, discarding inter-sentence dependencies, and existing benchmarks omit the newest generation of generators. We construct MOSAIC, a benchmark of 16,000 hybrid documents over PubMed and XSum, generated by DeepSeek-V3.2 and Kimi K2 under stringent quality controls including a perplexity-consistency filter absent from prior benchmarks. We recast S-AGTD as structured prediction over the document sentence sequence and instantiate it as SenFlow, integrating graph-based inter-sentence propagation with linear-chain CRF decoding in a single document-level pass over a sentence graph. SenFlow reaches state-of-the-art performance on MOSAIC, with a +4.15 pp average Macro-F1 margin on cross-domain transfer, the hardest of three protocols of increasing difficulty. We further find that even after the perplexity filter equalizes overt cues, AI insertions retain a generator-dependent sentence-length gap that sentence-level detectors still exploit. Code and data: https://github.com/luojingkun22/SenFlow
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。