honeybadger
源码formal VM benchmark and inspectable reasoning runtime for testing whether language models can follow machine-like execution semantics on synthetic tasks.
formal VM benchmark and inspectable reasoning runtime for testing whether language models can follow machine-like execution semantics on synthetic tasks.