Who's Behind the Text? Robust Model Authorship Attribution
作者: Minseok Ko, Hyunseok Lee, Woomin Song, Beomjun Kim, Jinwoo Shin, Kunwoong Kim
分类: cs.AI, cs.CL, cs.LG
发布日期: 2026-10-14
💡 一句话要点
提出ATLAS以解决模型作者归属问题
🎯 匹配领域: 支柱二：RL算法与架构 (RL & Architecture) 支柱九：具身大模型 (Embodied Foundation Models)
关键词: 模型作者归属 大型语言模型 文本生成 鲁棒性 源条件建模 学术诚信 网络安全
📋 核心要点
- 现有模型作者归属方法在面对未见领域和版本变化时，性能显著下降，无法有效应对分布转移问题。
- 本文提出的ATLAS通过源条件文本建模，比较文本在不同源下的生成可能性，从而实现更可靠的作者归属。
- 实验结果显示，ATLAS在多个基准测试中表现优异，尤其在复杂的生成条件下，准确率显著高于现有方法。
📝 摘要（中文）
随着大型语言模型（LLMs）在文本生成中的广泛应用，其滥用已成为日益严重的社会问题，涉及学术不端、模型蒸馏和网络攻击等方面。因此，模型作者归属（MAA）变得尤为重要，以促进LLMs的负责任使用并提高AI安全性。现有MAA方法在面对未见领域和写作环境时，性能往往显著下降。为此，本文提出了基于源条件文本建模的作者追踪方法（ATLAS），通过比较文本在各源条件下的生成可能性来进行归属。实验结果表明，ATLAS在MAABench和OpenTuringBench上均优于现有基线，尤其在生成领域和模型版本变化时，准确率达到74.27%，比最强基线高出9.00个百分点。
🔬 方法详解
问题定义：本文旨在解决模型作者归属（MAA）问题，现有方法在面对未见领域和版本变化时，性能显著下降，无法适应分布转移的挑战。
核心思路：ATLAS通过源条件文本建模，学习每个源家族的文本生成方式，并通过比较文本在各源条件下的生成可能性进行归属，避免直接预测源的局限性。
技术框架：ATLAS的整体架构包括源条件建模、文本生成可能性计算和基于token的证据聚合。通过对每个token的可能性进行剪裁，确保没有单一token主导决策。
关键创新：ATLAS的主要创新在于采用源条件建模而非直接源预测，这一设计使得模型在面对未见数据时依然保持较高的准确性，显著提升了MAA的鲁棒性。
关键设计：ATLAS在参数设置上进行了优化，采用了适应性损失函数和多层次的网络结构，以增强模型对不同文本生成条件的适应能力。
🖼️ 关键图片
📊 实验亮点
ATLAS在MAABench和OpenTuringBench上的实验结果显示，准确率达到74.27%，比最强基线高出9.00个百分点，展现了其在复杂生成条件下的优越性能，尤其在生成领域和模型版本变化时表现突出。
🎯 应用场景
该研究的潜在应用领域包括学术诚信监测、网络安全防护和模型蒸馏分析等。通过有效识别文本来源，ATLAS能够帮助机构和个人防范不当使用大型语言模型，提升AI系统的安全性和可信度。未来，该方法可能在更广泛的文本生成和分析任务中发挥重要作用。
📄 摘要（原文）
As large language models (LLMs) are increasingly used to generate text, their misuse has become a growing societal problem, with implications for academic misconduct, model distillation, and cyberattacks. For these reasons, Model Authorship Attribution (MAA), which identifies the source model of a given text, is important for promoting the responsible use of LLMs and improving AI safety. In practice, attribution systems must remain reliable beyond the distributions observed during training. They may encounter texts from unseen domains and writing settings, as well as generations from future versions of model families that were not available when the system was trained. However, existing MAA methods often exhibit substantial performance degradation under such distribution shifts. To address this challenge, we propose Authorship Tracing via Likelihood-based Attribution of Sources (ATLAS), which formulates MAA through source-conditioned text modeling rather than direct source prediction. ATLAS learns how text from each source family is generated and attributes a given text by comparing its likelihood under each learned source condition. We aggregate this likelihood evidence at the token level with clipping so that no single token dominates the decision. We further introduce MAABench, a broad MAA benchmark for evaluating practical OOD settings such as domain and version. Our experiments show that ATLAS consistently outperforms existing baselines across MAABench and OpenTuringBench. In particular, when both the generation domain and the model version change on MAABench, ATLAS achieves 74.27% accuracy, 9.00pp higher than the strongest baseline. Beyond MAA, ATLAS can also help identify distillation source models by measuring which source model a target LLM most closely resembles. Our extended experiments show that ATLAS recovers the known teacher of every distilled model we evaluate.