Implicit Causality-biases in humans and LLMs as a tool for benchmarking LLM discourse capabilities

Open in new window