Thunder-NUBench

Thunder-NUBench (Negation Understanding Benchmark) is a benchmark specifically designed to evaluate large language models’ (LLMs) sentence-level understanding of negation. Thunder-NUBench introduces rich, manually curated sentence pairs and multiple-choice tasks that contrast standard negation with structurally similar distractors (e.g., local negation, contradiction, paraphrase). The goal is to probe semantic-level understanding of negation.