学术前沿

DoReMi: Grounding Language Model by Detecting and Recovering from Plan-Execution Misalignment

提出即时检测与即时恢复框架。