Benchmark
OpenBookQA
An open-book question-answering benchmark dataset that simulates an open-book examination setting, requiring models to combine core scientific knowledge with commonsense reasoning. It is presented in a multiple-choice format.
Timeline (last 90 days)
Sep 18
VisKG-LM achieves improvements of 1.2, 0.8, and 4.3 points over the GreaseLM baseline on the three benchmarks CommonsenseQA, OpenBookQA, and MedQA-USMLE, respectively.
Unverified50%