Last year, I presented AutoFR, a reinforcement learning framework that automatically generates filter rules for adblocking optimized per-site. This year, I extend this work and provide a comparative analysis of various new methodologies to generate filter rules that can generalize across multiple sites.