Useful Diversity or Hallucinated Variety? Persona-Conditioned Synthetic Query Generation for E-Commerce Search

arXiv:2610.35102v1 Announce Type: cross Abstract: Synthetic supervision can bootstrap semantic parsers in new search domains, but prompt-induced diversity is useful only when generations are faithful to the structured intent. We explore this for cold-start e-commerce search: 118,098 synthetic pairs produced by 6 LLMs with 9 persona-conditioned prompt families on a 47-field UK automotive schema, against 1,214 human-annotated and 15,304 production queries. We show that the IPIP personality bundle improves parser F1 in all four matched contrasts (Holm-corrected $p<.001 and all three bundles have a positive main effect over the complete grid. syntax conditioning is axis that moves generations towards real traffic. it also induces lexical diversity. alignment skl divergence between von mises-fisher distributions fitted to synthetic human-annotated queries an unlabelled proxy for parser f1 against human written simulated shopper scenarios lower closer llm committee checked blinded author pass triages missed by every labels as hallucinated. dropped constraints dominate. four syntax-bearing families largest hallucinated shares while ipip alone has lowest at>