Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
Hefei University of Technology, Hefei, China
2
0
2
LA-RL achieves up to 20 F1 points improvement in out-of-distribution relation extraction by aligning self-reflection with task-specific error diagnostics.
Incorporating negative examples into in-context learning boosts LLM performance in information extraction by revealing the underlying causes of errors.