Understand this story
Google DeepMind and UPenn published a paper proposing an interpretable AI text detector that explains its decisions using readable writing features while achieving accuracy close to black-box detectors. The research trains a sparse autoencoder on the internal activations of a Gemma-2-2B detector, decomposes its activations into features, and has an LLM label each feature, retaining only features whose labels can reasonably predict their behavior as input to a simple classifier. Using only well-labeled features, the detector achieves an in-domain F1 of 96.8, close to the black box's 98.0; on unseen test data, the version based on base model features achieves an F1 of 92.2 versus 75.2 for the black box. The work is currently released as a paper with no further progress yet.
AI-generated from public reports
Latest development · Report
Google DeepMind and UPenn paper proposes an interpretable AI text detectorDevelopments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 23
Rohan Paul
Google DeepMind and UPenn paper proposes an interpretable AI text detector
A new paper from Google DeepMind and UPenn proposes an interpretable AI text detector that explains its decisions using readable writing features while achieving accuracy close to black-box detectors. The research trains sparse autoencoders on the internal activations of a Gemma-2-2B detector, decomposes its activations into features, and has an LLM label each feature, keeping only features whose labels can reasonably predict their behavior as input to a simple classifier. Using only well-labeled features, the detector achieves an in-domain F1 of 96.8, close to the black box's 98.0; on unseen test data, the version based on base-model features achieves an F1 of 92.2 versus 75.2 for the black box.