📝 New preprint out on on the limits of LLMs in lexical entailment recognition 🔬