Open-Ended Questions
Benchmarks
TaskDatasetSOTA resultTrendResults
Open-ended questions
68.9Win Rate
6
Open-Ended Questions (Human Generated Questions)
70.1Instruction Following Rate
2
Open-Ended Questions GPT-4 Generated Questions
96.9Instruction Following Rate
2
Open-Ended Questions Generated Questions GPT-3.5-Turbo
95.9Instruction Following Rate
2