BenchmarkBIG-Bench / Beyond the Imitation Game Benchmark
BigBench
A large-scale collaborative benchmark initiated by Google, comprising more than 200 diverse tasks. It is designed to evaluate reasoning and knowledge tasks that exceed the capability boundaries of current language models.
Timeline (last 90 days)
Sep 30
Mainstream large model benchmarks such as MMLU, BigBench, and MATH have all been widely used in training data filtering and reinforcement learning reward signal design, making it increasingly difficult for high scores to map to real capability improvements.
Unverified50%