Search: “routing”

arXiv:2609.15193v1 Announce Type: new Abstract: Drifting models offer a promising route to faster generative AI: they perform gradual transport during training, while generating new samples in a singl

arXiv:2609.12890v1 Announce Type: new Abstract: In autoregressive forecasting, long prediction rollouts provide distant supervision, but backpropagation through time (BPTT) carries gradients from thos

arXiv:2609.12686v1 Announce Type: cross Abstract: Large language models (LLMs) process long contexts, including long documents and lengthy conversations, but face token-level memory usage that increas

arXiv:2609.10964v1 Announce Type: new Abstract: Agentic LLM workflows consist of sequences of model turns interleaved with tool interactions, so their end-to-end completion time depends not only on in

arXiv:2609.11192v1 Announce Type: new Abstract: Soft context compression condenses a context into a few memory tokens that a frozen LLM consumes in place of the raw text, but existing compressors fix

arXiv:2609.08999v1 Announce Type: new Abstract: Partially Relevant Video Retrieval (PRVR) retrieves untrimmed videos when queries describe only short moments. Although recent methods improve local rep

arXiv:2609.05820v1 Announce Type: new Abstract: We study adaptive routing of prompts to large language model (LLM) experts to maximize response quality in an online setting with limited feedback. We f

Executive Summary Since the release of our May 2026 report detailing adversarial misuse of artificial intelligence (AI), Google Threat Intelligence Group (GTIG) has observed forward leaning adversarie

arXiv:2609.04453v1 Announce Type: cross Abstract: Expert pruning reduces the memory and serving cost of Mixture-of-Experts (MoE) models by removing low-importance experts identified by the router, ass

arXiv:2609.02928v1 Announce Type: cross Abstract: Modern multi-agent systems are increasingly deployed at scale over large populations of agents in settings such as ad-auctions, traffic routing, and r

arXiv:2609.02881v1 Announce Type: new Abstract: We introduce the Graph Machine (GM), an architecture that maintains an O(n)-sized state and accesses it through sparse, dynamic routing. Unlike methods

arXiv:2609.02293v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) is a scaling architecture for large language models that activates only a small subset of expert modules per token, enabling

arXiv:2609.00092v1 Announce Type: new Abstract: Long-horizon complex tasks require foundation models to accumulate information, maintain internal states, and adapt over extended interactions. Safety s

arXiv:2609.00597v1 Announce Type: new Abstract: With the rapid rise of large language models (LLMs), controlling undesirable model behaviors has become increasingly important. Existing behavioral cont

arXiv:2608.29088v1 Announce Type: new Abstract: Multimodal question answering remains sensitive to noisy, incomplete, and weakly grounded evidence. Long unstructured contexts can introduce redundancy

arXiv:2608.25992v2 Announce Type: replace Abstract: Multi-agent large language model (LLM) workflows have emerged as a powerful paradigm for solving complex, open-ended tasks through collaborative rea

arXiv:2608.28142v1 Announce Type: new Abstract: Electrification of commercial delivery fleets is shifting fleet routing from distance- and time-based optimization toward energy-aware decision-making.

arXiv:2608.28327v1 Announce Type: cross Abstract: Practitioners defend large language models (LLMs) by stacking defenses, assuming the layers compound. A stack is an ensemble, and ensembles compound o

arXiv:2510.09316v3 Announce Type: replace-cross Abstract: We compile 129 public LLM prompt datasets with more than 1.22TB and more than 673M instances and organize them into a unified taxonomy. We use

arXiv:2608.27038v1 Announce Type: new Abstract: Although batch prompting makes large language model inference more efficient by processing multiple instances simultaneously, it suffers from unpredicta

arXiv:2608.21909v1 Announce Type: new Abstract: While Multi-Task Learning (MTL) is essential for adapting Large Language Models (LLMs) to diverse domains, prevailing LoRA-based methods rely on complex

arXiv:2608.22820v1 Announce Type: cross Abstract: Deep learning models often produce performance disparities across demographic groups, due to the training data imbalance with respect to sensitive att

arXiv:2608.20757v1 Announce Type: cross Abstract: We describe the PSK submission to the WMT 2026 Multilingual Instruction Shared Task. Our system uses the 3.35B-parameter Tiny Aya Global model with th

arXiv:2608.20810v1 Announce Type: cross Abstract: Multimodal information systems increasingly route generated visual content back through the same vision-language index that informed its production, s

ChatGPT is usually treated as a prompt box. Ask for a cinematic shot, add a few camera words, copy the result into a video model, and hope the ending survives. My problem was that the ending usually d

arXiv:2608.17209v1 Announce Type: cross Abstract: End-to-end vision-language-action (VLA) and world-action models offer an elegant route to general-purpose robotics, but their reliability is bounded b

Real-time streaming pipelines are the operational backbone of modern enterprises, continuously processing everything from customer support interactions to transaction logs. Traditionally, streaming DA

arXiv:2608.16435v1 Announce Type: new Abstract: In collection operations, accumulating payload progressively slows the vehicle, imposing a cumulative penalty on routing efficiency. An onboard drone ca

arXiv:2608.15311v1 Announce Type: new Abstract: Federated instruction fine-tuning enables Large Language Models (LLMs) to adapt to decentralized, privacy-sensitive data without requiring data sharing.

Written by: Alex Tselevich, Michael Maturi Introduction Adversarial misuse of AI has increased the risk of data theft and extortion events, because when proprietary source code is exposed, defenders m

arXiv:2608.16502v1 Announce Type: new Abstract: Large-scale agents increasingly rely on retrieval to access external capabilities. We study this retrieval gate in structured tools and APIs, a measurab

arXiv:2608.14531v1 Announce Type: new Abstract: Articulated vehicles are the workhorses of freight transportation, and their autonomous navigation is challenging. Their physical characteristics and mo

arXiv:2608.11218v1 Announce Type: new Abstract: Independent agents that reason in latent space can share computed state as key-value cache fragments rather than text. Merged by a conflict-free replica

arXiv:2608.08402v1 Announce Type: new Abstract: Fiber-bundle endoscopy offers a compact and flexible route for clinical fluorescence imaging through natural human orifices, but since its first report

I wanted to find out whether a huge text-only MoE could be given basic vision without retraining the language model itself. The short answer is yes. I froze DeepSeek V4 Flash and a 417M-parameter Moon

arXiv:2608.08700v1 Announce Type: new Abstract: Reliable evaluation of tool routing is critical as Large Language Models increasingly operate as autonomous agents. Current benchmarks face three struct

arXiv:2608.08627v1 Announce Type: new Abstract: Sparse mixture-of-experts (MoE) layers expand recommendation capacity through conditional computation, yet a trained checkpoint still stores and routes

arXiv:2608.07994v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) is essential for enterprise knowledge question answering (QA), particularly in domains with complex product documen

arXiv:2608.08453v1 Announce Type: new Abstract: Under the current standard, Agent Skills are SKILL.md files that combine instructions with supporting resources, enabling Large Language Model (LLM) age

arXiv:2608.00591v2 Announce Type: replace Abstract: A calibrated stochastic world model can reveal how uncertain a future is without revealing why it branches. The same conditional future law can aris

arXiv:2608.06516v1 Announce Type: cross Abstract: Lightweight connectors make frozen multimodal encoders composable at the representation level. Deployment exposes a second problem at the level of tas

TensorSharp's MoE CPU-offload feature has been merged into main. Here is the parameters description of this feature: Mixture-of-Experts CPU offload: --n-cpu-moe <N> | -ncmoe <N> Keep the routed MoE ex

arXiv:2608.01792v1 Announce Type: cross Abstract: Intelligent document processing (IDP) with vision-language models (VLMs) hinges on confidence scores trustworthy enough to route extractions between a

arXiv:2607.28679v1 Announce Type: new Abstract: Multi-agent planning problems arise in a variety of engineering applications, such as multi-robot wildfire fighting and unmanned aerial inspection in fa

A few people asked for this after the bandwidth thread, so here it is on its own instead of buried in a comment. The dense rule was simple: every token reads every weight, so tokens/sec ≈ bandwidth ÷

arXiv:2607.24797v1 Announce Type: cross Abstract: In the literate human brain, reading and writing are two doubly-dissociable systems: a ventral decoding route (impaired in pure alexia) and a fronto-p