Metraipolis › archive › dc_63701502bdee
When Learned Context Planning Fails to Beat Strong Retrieval: A Controlled Study of Planning, Routing, and Reranking for Long-Context QA — arXiv:2609.26976
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.26976
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T10:04:42.519Z
- SHA-256 of the archived text
1269650dd94d6c01d93736c660e7c61d3fbc4045d28d7a37055bd5a5041cce78- Stable URL of this page
https://metraipolis.com/documento/dc_63701502bdee
How to cite
When Learned Context Planning Fails to Beat Strong Retrieval: A Controlled Study of Planning, Routing, and Reranking for Long-Context QA — arXiv:2609.26976 — arXiv (en), https://arxiv.org/abs/2609.26976 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_63701502bdee (Metraipolis archive dc_63701502bdee, sha256 1269650dd94d6c01d93736c660e7c61d3fbc4045d28d7a37055bd5a5041cce78)
BibTeX
@misc{dc63701502bdee,
title = {When Learned Context Planning Fails to Beat Strong Retrieval: A Controlled Study of Planning, Routing, and Reranking for Long-Context QA — arXiv:2609.26976},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.26976},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_63701502bdee; sha256 1269650dd94d6c01d93736c660e7c61d3fbc4045d28d7a37055bd5a5041cce78},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.26976. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Learned context planning selects evidence atoms before an answer model reasons over them. We test whether this learned selection improves long-context multiple-choice QA after strong retrieval, routing, budgeted-selector, and reranking controls. Our primary diagnostic uses all 503 LongBench-v2 MCQ questions with Qwen2.5-7B-Instruct. The planner is SFT-trained on outcome-selected traces from 140 training and 28 development questions; because the 503-question analysis includes those questions, it is partly transductive. At an 18k-character budget, anchored hybrid retrieval reaches 36.18% accuracy and BM25 reaches 35.98%, while the best direct planner-guided method reaches 34.19%. On the untouched 152-question test split, anchored hybrid remains higher (42.11% versus 36.84%). Leakage-safe routers cannot convert a large oracle gap. Under tight budgets, the best planner is ahead by only 0.40 points at 6k and loses at 9k; planner-guided reranking has a +1.79-point estimate at 6k with a paired interval crossing zero and ties the control at 9k. Packing-order and score-flatness analyses did not identify a stable mechanism. Under this setup, learned planning is a weak relevance signal rather than a replacement for strong retrieval. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-24 in cs.CL). Categorie arXiv: Computation and Language (cs.CL) ; Information Retrieval (cs.IR). Pagina del lavoro: https://arxiv.org/abs/2609.26976 — 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 2026-10-08T09:11:49.582Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-24?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).