Beyond Easy Wins: A Text Hardness-Aware Benchmark for LLM-generated Text Detection

Open in new window