Metraipolis › archive › dc_4494d7ead85d

MERGE: Multi-LLM Ensemble for Retrieval via Generative Enrichment — arXiv:2609.37574

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.37574
License (as archived)
CC0-1.0
Archived
2026-10-07T16:57:40.087Z
SHA-256 of the archived text
98309e861873b975901e7577859f6685c9b5224bd3ba7da233d9f4f5fae195e7
Stable URL of this page
https://metraipolis.com/documento/dc_4494d7ead85d

How to cite

MERGE: Multi-LLM Ensemble for Retrieval via Generative Enrichment — arXiv:2609.37574 — arXiv (en), https://arxiv.org/abs/2609.37574 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4494d7ead85d (Metraipolis archive dc_4494d7ead85d, sha256 98309e861873b975901e7577859f6685c9b5224bd3ba7da233d9f4f5fae195e7)
BibTeX
@misc{dc4494d7ead85d,
  title = {MERGE: Multi-LLM Ensemble for Retrieval via Generative Enrichment — arXiv:2609.37574},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.37574},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4494d7ead85d; sha256 98309e861873b975901e7577859f6685c9b5224bd3ba7da233d9f4f5fae195e7},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.37574. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large Language Models (LLMs) are increasingly used to enrich user queries in information retrieval (IR) so that a standard retriever such as BM25 can bridge vocabulary gaps with the target corpus. Any single LLM, however, is limited by its training data and architectural biases, and its enrichment behavior depends on hand-crafted prompts that must be re-engineered for each new model -- an expensive and poorly scalable process. We present MERGE (Multi-LLM Ensemble for Retrieval via Generative Enrichment), a two-stage framework: three heterogeneous 7-8B open-source LLMs independently produce candidate expansions, and a larger LLM generatively synthesizes them into a single query. To make prompt engineering scalable across the ensemble, we integrate a task-grounded Automatic Prompt Optimization (APO) loop into both stages. Unlike APO methods that judge candidates with an LLM evaluator, our loop scores each candidate by its downstream retrieval performance and runs a small tournament between the current champion prompt and optimizer-proposed drafts, terminating once the champion survives two consecutive rounds; a history-augmented variant additionally feeds the recent tournament trajectory back to the optimizer. MERGE is retriever-agnostic and issues a single BM25 pass with no rank fusion, no supervised document expansion, and no re-indexing. On five BEIR benchmarks (NQ, SciFact, FiQA, Touche-2020, DBPedia), MERGE improves BM25 nDCG@10 over the original queries by +2.1 to +14.9 points and matches or outperforms strong LLM-based query-expansion baselines despite using only compact open-source models. Ablations confirm that the Stage-2 ensemble beats any single Stage-1 LLM, and that task-grounded APO converts large seed-prompt regressions into consistent gains without hand-tuning.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.CL).
Categorie arXiv: Information Retrieval (cs.IR) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.37574 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T16:27:03.450Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).