List your company

ResearchFormerly WizWand

Head-QA

Benchmarks

TaskDatasetSOTA resultTrendResults

Reinforcement Learning from Verifiable Rewards

HEAD-QA

100AR

30

Medical Multiple-Choice Question Answering

Head-QA (test)

28.79Accuracy

10

Question Answering

HEAD-QA

100AR (%)

7

Question Answering

HEAD-QA (eval)

0Pathwise Violations (PathV)

3