NishiSearch

Web This site Trusted

30 result(s) for “okapi bm25 site hotchips pp site can kernel org site www amet cts com site:arxiv.org” of 33 matched · 640 ms

Scope: Nishi’s own index (our crawl + Common Crawl) — not the whole live web. Some sites block all automated crawlers (anti-bot walls), so their pages may be absent here even though we never filter lawful content.

Freshness: no result on this page carries a fetch time yet (indexed before the fetch epoch was stored beside the doc)

The query asks the model to recall the topic of an earlier conversation about Tesla; the gold answer is anchored in one specific session about Pooja’s dig site
from the top result below

Report GitHub Issue

https://arxiv.org/html/2608.02613v1
We evaluate five open-weight readers with Vanilla context, BM25-RAG, Oracle retrieval, Memobase, and MemSearch as memory backends

%PDF-1.7 % 1 0 obj > stream xYr8}c{R$ʑ׎l#9KMuh

https://arxiv.org/pdf/2608.02122
) /Title (DEFT: Joint Task Placement and DVFS for Energy-Efficient Multi-GPU Runtimes) /Trapped /False /arXivID (https://arxiv

1 Introduction

https://arxiv.org/html/2501.13956v1
com Abstract We introduce Zep, a novel memory layer service for AI agents that outperforms the current state-of-the-art system, MemGPT, in the Deep Memory Retrieval (DMR) benchmark

Report GitHub Issue

https://arxiv.org/html/2608.03048v1
for large language models, as recent recurrent-memory approaches face two inherent challenges: sequential chunk-wise updates can overwrite early critical evidence with later irrelevant content, and serial inter-chunk dependencies limit parallelism and cause latency to increase with context length

Artificial Intelligence

https://arxiv.org/list/cs.AI/new
We demonstrate that it can self-assemble functional circuits from scratch and rapidly re-route logic around permanent, previously unseen hardware faults

%PDF-1.7 % 1 0 obj > stream xZYS9~_ݘ0*tUĄ#pc0c6Mz

https://arxiv.org/pdf/2608.03118
Zz|PG-T*1P8Yvk-'Nğ`PM$&] L'|*p7k7mE7nry09U=PP#'#ބv2Z$t;Ʀlr5

%PDF-1.7 % 1 0 obj > stream x\rFr}WQ`IihʫRZ

https://arxiv.org/pdf/2608.03411
6applМe~-y2Gi&pf3KjGKe ,"&:w|ҶP"rfr\xqF=/ܖn*>]zJX-w_0}0(B5&2

%PDF-1.7 % 1 0 obj > stream x}Qn0+Ǝ_ BR>NC,)$1}

https://arxiv.org/pdf/2607.27070v1
IRGsro>tJ~~z|AOρ{U+(>[c;JA[/0Ҁg^9xӁwVo9ѳICÁ`xG]򩾥WW6W'vѴ>/+ڀmƷޝE{11E-Lw6 ܿ]_qеDuykG $AbpPЍkZ_E8JVkj&}0*V: HH

%PDF-1.5 % 1 0 obj > endobj 2 0 obj > endobj 3 0 obj > stream Event

https://arxiv.org/pdf/2508.07408
-Aware Sentiment Factors from LLM-Augmented Financial Tweets: A Transparent Framework for Interpretable Quant TradingYueyi WangQiyao Weihttp://creativecommons.org/licenses/by/4.0/q-fin.STcs.CLcs.LG endstream endobj 4 0 obj > endobj 5 0 obj

 ^TU_/ J.twKtw`ҠHwHt H(R}\

https://arxiv.org/src/2006.15704
wν~cƚ\s5fVz&F\@ ;߯'i qsr >gq?z_QFHϊDZp(@(pn}"Ϩ(*0Mrq@ F0 Ru5p9Yژq "~ (h~$nc|O+9:

%PDF-1.7 % 1 0 obj >/MarkInfo>/Metadata 2783 0 R/ViewerPreferences

https://arxiv.org/pdf/2608.02976
2784 0 R>> endobj 2 0 obj > endobj 3 0 obj >/ExtGState>/ProcSet[/PDF/Text/ImageB/ImageC/ImageI] >>/Annots[ 18 0 R] /MediaBox[ 0 0 612 792] /Contents 4 0 R/Group>/Tabs/S/StructParents 0>> endobj 4 0 obj > stream x}[sDCnµ"F

rj\nGvz`fLRel.][:{kfwuXgm~l:

https://arxiv.org/src/2608.03966
N`dVh[OޤRzk")aJl)o`)IϏH҇tX)M\6"PPPZj;*9sBu Id\qVsI012(nTOle_c&B h FD$

4^h?Gsz:S;:d|2ǣOÓOR%['

https://arxiv.org/src/2006.00508
6t[Þ9{ʞ{Wwb {ޱ #;X/ŭIf30q6g d]3,dE=YRvVl>H:OAibvĆlĎ٘Sv ;!"H )5Naxzj8џ74y jHx:_yf!~Noi2VoZ

k1a}kwh^ \&>璴޶L6ÖlIJI L?$S6">LDY2et"ڌG

https://arxiv.org/src/2104.08691
>6JfHx@OFm1Id4L$3~8q=vtz3D w9Hpր `uy[z[(VD8\I,R/(~ǺzYSΫӑQ #SgKI!p 8 , #X6Lp;8@$"ZLU67^B+h!CT[-8y# E;ܟdA!\O}ܟ(

%PDF-1.7 % 1 0 obj > stream x[[s8}_Ky U[[[x$㌓y%Zf

https://arxiv.org/pdf/2608.02867
,I9q~w%"i{I4Nn4ANU}^.>i>-?\NiёD) o~דoʊ5j_(𯏏rY@*%]S-is[#_ Ąn ]80 IFpwpwV/.q XËˣlRˌy8oD(OxiӚ,M

%PDF-1.3 % 3 0 obj > stream x\rIv}W1*vTElͱ&;

https://arxiv.org/pdf/2407.01722
y( Y hܛKPT7A[DeVͻoOo`(N44 +,(qEEl߃V2Fx`t|Di8e!#I88/(4Me[DMo$`> \s@: $HɻW?, +a~s``8*M5?k`Cyߚ9`

^񰏽 NkzJrYY^hBB۷'ucRʲ

https://arxiv.org/src/2006.01331
>'"J=&_[:$@0՛^Ajj> [AQt`}~Q?z(Rx{>2J`һA;2?;_ڱ퀂.OVcбn{¾c){ƞ_ oKf?7o-{.؏'3c&͘\ٜ-ؒy7v| ,f KYn-[5 Cdj"F?&8C%(E|C$

_uTe' Cakaw( /p{uνt;otxo?

https://arxiv.org/src/2006.03594
6gb3=};9jB^MM=qL^pp>uNd:En=ʢ(_o饖X,(RcP9#4ٜ+R0ʰǬe{'{";(:[mjqU8y>erA( ;]V5Xclc}1eڊ06ȯ~Ps hGyqJ

W>h[msFg)T$lG^+已]  H@ >=3

https://arxiv.org/src/2506.01844
(Rvv۪3~ɋޞxo!gp?npQJ݃ f=ur{ҔPWxN ,_z=HNZ$Y0 ɼZ3V; &,/vrFagW7z"gWҝY4:|ԓy`"ņfA퍫+ k?a:vYR-"z )0A5

Computation and Language

https://arxiv.org/list/cs.CL/new
LG) Half a million cuneiform clay tablets survive in museums worldwide, yet modern users can neither read nor write in the world's oldest writing system, leaving a 4,000-year cultural barrier that existing NLP tools have only partially addressed

[2608.03860] SciRet: A Compute-Aware Empirical Study of Retrieval and

https://arxiv.org/abs/2608.03860
The pipeline combines sentence-window chunking, BM25, BGE-M3 dense retrieval, reciprocal rank fusion, optional cross-encoder reranking, and grounded answer generation

1 Introduction

https://arxiv.org/html/2407.16741v3
com/All-Hands-AI/OpenHands Slack http://bit

%PDF-1.7 % 1 0 obj > stream x3R0P0V02䲱QwI-.)V040 )ف}SKSK

https://arxiv.org/pdf/2607.26245v1
f`q}b#ǿ 51$3?O*VZTcaHLOOIU-N˂-!B* V䧃ld~+ endstream endobj 4 0 obj > endobj 5 0 obj > stream xZis8ʀ8HUSS#d&grm%'P̯$(9[Gׯa( 0H򀇁灌Ӏ@)4 Ѩ8Dc

Bandwidth-Free Inference for Recursive Nonlinear Impulse Response

https://arxiv.org/html/2608.02943v1
org perpetual non-exclusive license arXiv:2608

[2608.03883] DS@GT-ARC at eRisk 2026 Task 3: Sparse, Semantic, and LLM

https://arxiv.org/abs/2608.03883
Our systems combine sparse BM25 retrieval, evidence-aware rescoring for self-referential symptom reports, embedding-based reranking, query-prototype expansion, and LLM-based reranking

Information Retrieval

https://arxiv.org/list/cs.IR/recent
IR) Tue, 4 Aug 2026 (showing first 34 of 45 entries ) [17] arXiv:2608

[2104.08663] BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of

https://arxiv.org/abs/2104.08663
Our results show BM25 is a robust baseline and re-ranking and late-interaction-based models on average achieve the best zero-shot performances, however, at high computational costs

Information Retrieval

https://arxiv.org/list/cs.IR/new
Our results show that skill representation itself matters, and that simply preserving the structure already present in skill files can substantially improve retrieval

cCc\{s6ϿO؛DVD\bgܤ&۫Υ +|Vx $He$

https://arxiv.org/src/2210.03493
m>N#0YU2z͠]7pP+WU˕ytR~8:Kf-g`8χnupn~Z(=@튑0܈m^o;1xQi\5 |#)+J/@n Osa

%PDF-1.7 % 1 0 obj > stream xiSѮG**&ر dUd

https://arxiv.org/pdf/2607.26327v3
'-~}HZ$ {-0B/pF '|H‹BI-EJ+-0$sJ( #,5BKZb' 0vA,:i82 #Mh0i-aMB[aVF0K 3J8F8 t 9 2a5Qa}.!PIJp^J 6;Z#Nea>6EpYJ