Scoring Explanations Without Reading Them: Structural Shortcuts in Automated-Interpretability Evaluation Data
arXiv:2610.37676v1 Announce Type: new Abstract: Automated interpretability scores an explanation of a sparse-autoencoder (SAE) feature by asking a model to predict, token by token, where the feature fires on evaluation examples or to judge whether marked tokens fit it. We find that the evaluation da
Der Artikeltext wird beim nächsten Sync nachgeladen. Original bei arXiv cs.CL lesen
Quelle: arXiv cs.CL (arXiv cs.CL). Original öffnen