- gpu_experience_collector: branch_sizes [5,3,3]→[9,3,3] - gpu_iqn_head: branch_0_size 5→9 - All DQN configs: num_actions 5→9 - Production TOML: branch_0_size=9 - Removed use_branching from TOMLs (always enabled) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>