NormDetect: Anchored Calibration for Low-False-Positive Detection of Machine-Generated Web Text

Summary

NormDetect is a detector for machine-generated web text that anchors a learned per-section head to the normalized Binoculars zero-shot score, using the Falcon-7B scoring pair. On eight sections of the RAID benchmark with eleven attacks, it raises recall at a 0.1% calibrated false positive rate by 5.0 points over normalized Binoculars, while keeping the human false positive bound at most 1% in every section.

Related items

• InfoVulnerabilities and Defenses in Audio-Visual Attacks: A Survey From Audio to Multimodal ModelsSimilar attack · ACM Digital Library (TOPS, DTRAP, CSUR)
• InfoDesign of an Iterative Adversarial Resilient Graph Neural Networks for Intrusion Detection in Software Defined Networks and DeploymentsSimilar attack · OpenAlex (peer-reviewed AI security)
• InfoTextual Response-Aware Guidance for Zero-Shot Adversarial Robustness of Vision-Language ModelsSimilar attack · Arxiv (cs.RO + cs.CV security)
• InfoLiDAR-SHIELD: A Fully Dockerized Adversarial Robustness Assessment Framework for LiDAR-Based 3D Perception ModelsSimilar attack · Arxiv (cs.RO + cs.CV security)
• InfoPASA-KV: Attacking Action-Facing Visual Routes in World Action ModelsSimilar attack · Arxiv (cs.RO + cs.CV security)