List your company

Yes/No Question Answering on BoolQ (test)

79.2Accuracy4o
4856.164.272.3Jul 1, 2021Jan 28, 2022Aug 27, 2022Mar 26, 2023Oct 23, 2023May 21, 2024Dec 19, 2024

Evaluation results

MethodLinks
4oModel Family=ChatGPT2024.1279.24.911.3
o1-previewModel Family=ChatGPT2024.1278.74.913.2
R1Model Family=DeepSeek2024.1278.11.67.9
o1-miniModel Family=ChatGPT2024.1274.14.215.6
CLINEmode=fine-tuned2021.0773.9——
RoBERTamode=fine-tuned2021.0769.6——
V3Model Family=DeepSeek2024.12699.228.5
3.5-turboModel Family=ChatGPT2024.1262.512.134
BERTmode=fine-tuned2021.0760.9——
2-7BModel Family=Llama2024.1252.88.726.5
3-8BModel Family=Llama2024.1250.120.358.2
3.1-8BModel Family=Llama2024.1249.220.458.8