Investigating which interpretable language-model features predict headline click-through differences.
natural-language-processing xgboost ab-testing gemma sparse-autoencoders explainable-ai interpretable-machine-learning rulefit upworthy
-
Updated
Sep 24, 2026 - Jupyter Notebook