Metraipolis › archive › dc_71df8878f9cc

LODESTAR: Robust Entropy-Based Answer Selection in Retrieval-Augmented Generation for Question Answering -- Directing Frozen-LLM Entropy with a Reinforcement-Learned… — arXiv:2608.11922

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.11922
License (as archived)
CC0-1.0
Archived
2026-10-19T05:27:39.185Z
SHA-256 of the archived text
803b331949c17d3a76a4d98d1cf39dfbaf8c7d3cb2871150fdb873c07878a5eb
Stable URL of this page
https://metraipolis.com/documento/dc_71df8878f9cc

How to cite

LODESTAR: Robust Entropy-Based Answer Selection in Retrieval-Augmented Generation for Question Answering -- Directing Frozen-LLM Entropy with a Reinforcement-Learned… — arXiv:2608.11922 — arXiv (en), https://arxiv.org/abs/2608.11922 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_71df8878f9cc (Metraipolis archive dc_71df8878f9cc, sha256 803b331949c17d3a76a4d98d1cf39dfbaf8c7d3cb2871150fdb873c07878a5eb)
BibTeX
@misc{dc71df8878f9cc,
  title = {LODESTAR: Robust Entropy-Based Answer Selection in Retrieval-Augmented Generation for Question Answering -- Directing Frozen-LLM Entropy with a Reinforcement-Learned… — arXiv:2608.11922},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.11922},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_71df8878f9cc; sha256 803b331949c17d3a76a4d98d1cf39dfbaf8c7d3cb2871150fdb873c07878a5eb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.11922. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Predictive-distribution entropy is a strong answer-selection rule in retrieval-augmented generation (RAG) for question answering: across five QA benchmarks, selecting the answer a frozen respondent LLM produces with the lowest answer-token entropy lifts mean $F_1$ from 0.4769 to 0.5148 over the retriever's top-ranked passage, without gold answers. Yet this rule, which prior entropy-based selectors adopt, fails: a misleading passage makes the respondent confidently wrong, driving entropy down where the uncertainty signal looks most trustworthy. The failure comes from the passage the respondent reads, and the context it is read in is an input we can intervene on. We introduce LODESTAR: to our knowledge the first method to score a text intervention by the uncertainty it induces in a third-party frozen respondent, compared within one question. LODESTAR uses reinforcement learning (GRPO) to train, once and offline, a polarizer -- a short fixed natural-language string inserted into the respondent's prompt and never into its weights, directing entropy so that entropy-based answer selection stays robust to misleading passages; training labels are built from gold answers and two LLM judges, and inference reads neither. With every competing selector under the same frozen respondent and candidate pools on 5,008 questions, LODESTAR attains the highest mean $F_1$ of any inference-ready selector (0.5339), the highest macro exact match (0.4136), and the highest GPT-4o judge score of the frozen-respondent configurations judged (0.6435); its three-seed mean wins all 70 $F_1$ cells against fourteen published configurations and is paired-significant on $F_1$ against every one. The gain holds in-domain on NQ-Open and out-of-domain over SQuAD, TriviaQA, EntityQuestions and WebQuestions. Ablating the polarizer shows it is what makes the respondent read a misleading passage less often (26.0% vs 30.3%).

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-13 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Information Retrieval (cs.IR); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.11922 — 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 2026-10-19T05:15:39.224Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-13?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).