GroundCocoa: A Benchmark for Evaluating Compositional & Conditional Reasoning in Language Models

Harsh Kohli, Sachin Kumar, Huan Sun

The rapid progress of large language models (LLMs) has seen them excel and frequently surpass human performance on standard benchmarks. This has enabled many downstream application…