Most AI benchmarks test what systems can do in isolation; the Cooperation Benchmark tests how they behave in relationships. We evaluate whether AI systems support trust, cooperation, and mutual ben...
Açıklama
Most AI benchmarks test what systems can do in isolation; the Cooperation Benchmark tests how they behave in relationships. We evaluate whether AI systems support trust, cooperation, and mutual benefit across repeated interactions with people; whether they behave as a
