Last released Oct 16, 2026
Interpretable contextual-bandit learning from questions and action feedback