Improving Retrospective Language Agents via Joint Policy Gradient Optimization

Open in new window