Improving Open Information Extraction with Large Language Models: A Study on Demonstration Uncertainty

Benchmark Model Rank Results
open-information-extraction-on-carbGPT-3.5-Turbo w/ Selected Demo & UncertaintyF1: 52.1
open-information-extraction-on-carbLLaMA-2-13B w/ Selected Demo & UncertaintyF1: 36.2
open-information-extraction-on-carbLLaMA-2-70B w/ Selected Demo & UncertaintyF1: 51.5
open-information-extraction-on-oie2016GPT-3.5-Turbo w/ Selected Demo & UncertaintyF1: 65.1
open-information-extraction-on-oie2016LLaMA-2-13B w/ Selected Demo & UncertaintyF1: 36.9
open-information-extraction-on-oie2016LLaMA-2-70B w/ Selected Demo & UncertaintyF1: 65.8