Fetching the paper…

KOR-Bench: Benchmarking Language Models on Knowledge-Orthogonal Reasoning Tasks · Around