Tag
This paper proposes a multi-level contextual token relation modeling framework for machine-generated text detection, integrating local Markov-informed calibration and global rule-support reasoning to improve detection across cross-LLM and cross-domain settings with low computational overhead.
This paper addresses the degradation of likelihood-based machine-generated text detectors by identifying a Simpson's paradox in token-score aggregation. It proposes a learned local calibration step that significantly improves detection performance across various models and datasets.