Fetching the paper…

WorldSense: A Synthetic Benchmark for Grounded Reasoning in Large Language Models · Around