#Cypher

'AmbiGraph-Eval introduces a 560-example benchmark that measures LLM performance on ambiguous natural language to Cypher generation. Evaluations of nine models highlight challenges in ambiguity resolution and syntax generation.'