Metraipolis › archive › dc_8fdd5fee31dd
LENS: In-Context Search via Latent Evidence Exploration over Dynamic Raw Documents — arXiv:2608.16185
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.16185
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T18:07:03.064Z
- SHA-256 of the archived text
d7737b0eaf4edd56b1ec274490b92b5560dfd38d2642e3ae3788b2ee92db8304- Stable URL of this page
https://metraipolis.com/documento/dc_8fdd5fee31dd
How to cite
LENS: In-Context Search via Latent Evidence Exploration over Dynamic Raw Documents — arXiv:2608.16185 — arXiv (en), https://arxiv.org/abs/2608.16185 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_8fdd5fee31dd (Metraipolis archive dc_8fdd5fee31dd, sha256 d7737b0eaf4edd56b1ec274490b92b5560dfd38d2642e3ae3788b2ee92db8304)
BibTeX
@misc{dc8fdd5fee31dd,
title = {LENS: In-Context Search via Latent Evidence Exploration over Dynamic Raw Documents — arXiv:2608.16185},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.16185},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_8fdd5fee31dd; sha256 d7737b0eaf4edd56b1ec274490b92b5560dfd38d2642e3ae3788b2ee92db8304},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.16185. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
LLM agents increasingly answer questions over dynamic raw-document collections, where files may change before preprocessing, and relevant evidence (spans, sections, pages, or tables) is query-dependent. Existing retrieval-augmented approaches pre-materialize evidence via fixed chunking, embeddings, or persistent indexes: effective for lookup, yet costly, stale-prone, and committed to a granularity before the query is known. We formulate in-context search as Budgeted Evidence Localization over a latent evidence space induced by dynamic raw documents and propose LENS (Latent Evidence Exploration and Search), an index-free framework. Instead of pre-materializing the evidence space, LENS maintains a query-conditioned belief over candidate units, iteratively selecting candidates via complementary lexical, local, and exploratory proposal policies, updating the belief via an LLM relevance oracle, and narrowing toward high-posterior regions under a controllable budget. Evidence is consolidated into compact, source-grounded regions of interest and compressed into self-organizing knowledge clusters reused across related queries. On a controlled 500-question evaluation with matched corpus snapshots, LENS reaches 62.4% exact match and 84.8% evidence recall vs. 65.2% exact match but 50.4% evidence recall for a ReAct-style baseline. Across scales, LENS gives the strongest supporting-fact localization and answer grounding. On a fixed 150-question fullwiki subset over the raw Wikipedia dump with zero indexing, LENS and ReAct are nearly tied in official answer quality (43.3% vs. 42.7% EM), with LENS grounding more answers in retrieved evidence (84.0% vs. 70.7%). A no-retrieval Closed-Book reference highlights the contribution of model memory. LENS is query-ready after corpus changes, needs no preprocessing or persistent index, and preserves source-grounded evidence localization throughout. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-18 in cs.AI). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.16185 — 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 2026-10-18T16:29:14.091Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).