Benchmark
StrategyQA
A multi-hop commonsense reasoning benchmark that requires a model to combine multiple facts to perform reasoning.
Timeline (last 90 days)
Oct 22
Common reasoning benchmarks include GSM8K (elementary school math word problems), MATH (competition-level math problems), StrategyQA (multi-hop commonsense reasoning), and others; these benchmarks require multi-step derivation and have automatically verifiable answers.
Unverified50%