H. Le

QueryRoute is introduced, a benchmark that freezes the expensive artifacts needed to study this inference-time decision problem reproducibly: original queries, generated variants, ranked lists under multiple retrievers, retrieval scores, and per-query oracle labels.

EviQE is proposed, which aggregates documents retrieved by multiple reformulators, selects a compact evidence set, and uses it for one grounded expansion step, which separates evidence selection from generation and treats reformulators as complementary retrieval perspectives.

LLM-powered query reformulation, a hot topic in IR, often fails to translate gains from lexical to neural retrieval, and bigger models don't always help.