核心概念
Large Language Models can improve zero-shot information extraction by following annotation guidelines.
統計資料
Movie: 63
Restaurant: 21
Politics: 20
Literature: 31
Music: 24
AI: 41
Science: 41
GPT-3.5: SOTA
GoLLIE: Outperforms previous attempts
引述
"Large Language Models combined with instruction tuning have made significant progress when generalizing to unseen tasks."
"GoLLIE is able to improve zero-shot results on unseen IE tasks by being fine-tuned to comply with annotation guidelines."