Metraipolis › archive › dc_c6afb93d855a
Periscope: Extending Frozen Language Models Beyond Their Context Window — arXiv:2610.04047
- Source
- arXiv
- Original
- https://arxiv.org/abs/2610.04047
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T05:28:00.184Z
- SHA-256 of the archived text
8a91cd3ec86511f68df3cb48ce03a3073885635221e909960a6284197f0eae9c- Stable URL of this page
https://metraipolis.com/documento/dc_c6afb93d855a
How to cite
Periscope: Extending Frozen Language Models Beyond Their Context Window — arXiv:2610.04047 — arXiv, https://arxiv.org/abs/2610.04047 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c6afb93d855a (Metraipolis archive dc_c6afb93d855a, sha256 8a91cd3ec86511f68df3cb48ce03a3073885635221e909960a6284197f0eae9c)
BibTeX
@misc{dcc6afb93d855a,
title = {Periscope: Extending Frozen Language Models Beyond Their Context Window — arXiv:2610.04047},
howpublished = {arXiv},
url = {https://arxiv.org/abs/2610.04047},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c6afb93d855a; sha256 8a91cd3ec86511f68df3cb48ce03a3073885635221e909960a6284197f0eae9c},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.04047. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
A language model reads long text in one quadratic forward pass, stops at the context window, and loses accuracy with length before reaching it. We ask whether the read can be factorized when deciding over a finite set: which document is relevant, which option is supported, which passage is the evidence. Periscope, a training-free inference method, arranges the $N$ chunks of a text on a $K{\times}K$ grid with $K{=}\lceil\sqrt{N}\rceil$ and asks a frozen model the same question about $K$ local spans of consecutive chunks and $K$ strided spans that sample the whole text, reading the log-odds of every answer at one token. Each answer takes its best local and strided score, and scoring every chunk by its two spans gives an evidence map at no further cost, whose peak is the chunk behind the answer. Every probe is about $\sqrt{sc}$ tokens for a text of $s$ tokens and chunk size $c$, so a window of $W$ tokens reaches $W^{2}/c$ tokens at $s^{1.5}$ cost. The map replaces the long read. On LongBench v2, reading only the $K$ chunks the map ranks highest, 9k tokens, matches the same model's best window read across windows from 32k to 1M tokens, and on InfiniteBench, where the median context is 150k tokens, it leads the best window read by 5 points. The same map ranks BRIGHT's long-document corpora with the best NDCG@10 of six methods. Each call caches only one probe, so a 27B model reads 4.5M-token contexts on one 80GB GPU, where a single pass would need 296GB of cache. A long read then needs a GPU that holds the model, not one that holds the text.
Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Information Retrieval (cs.IR); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.04047 — 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 2026-10-06T05:24:20.245Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).