The paper introduces Voyager, a lightweight router that learns to select a sparse, final‑layer‑anchored expert from CLIP’s layer depth for out‑of‑distribution (OOD) detection. Using a weak OOD supervision context, Voyager outperforms state‑of‑the‑art prompt‑learning methods on three benchmarks, achieving an average FPR@95 of 18.86 on ImageNet‑1K. The router can be trained in about two minutes with less than 1 GB of memory, making it roughly 20× more efficient than existing approaches.
arXiv:2609.23600v1 Announce Type: new Abstract: Prompt learning efficiently adapts vision-language models (VLMs) to downstream tasks, but gains on seen classes often come at the expense of generaliza...
arXiv:2610.19198v1 Announce Type: cross Abstract: With the growing number of LLMs available, selecting which one to use for a prompt has become a key question. A large part of the routing literature...
The paper proposes a Learnware-based framework for deploying scene‑specific CSI feedback models in 6G systems. A centralized AI data center maintains a catalog of pre‑trained models, each tagged with semantic and statistical specifications. Base stations retrieve the most relevant model using only statistical fingerprints, which reduces data privacy risks, lowers retrieval latency, and cuts fine‑tuning effort, achieving up to 57.7% performance gains over a general model.
MetaNet is a support‑set controller that predicts, for each layer of a Mixture‑of‑Experts model, an expert‑retention threshold and a bounded routing bias while keeping the backbone, experts, and router frozen. On DeepSeek‑MoE‑16B‑Chat, MetaNet offers a tunable trade‑off between accuracy and expert activation: a conservative setting activates 3.61 experts on average (40% fewer than a fixed k=6) with comparable MMLU accuracy, whereas an aggressive setting activates only 2.28 experts (62% fewer) with a modest accuracy drop. The MMLU‑trained controller also transfers to C‑Eval, activating 2.90 experts on average (52% fewer than fixed k=6) at 0.386 accuracy.
arXiv:2604. 00421v2 Announce Type: replace Abstract: Mixture-of-Experts (MoE) layers increase model capacity by activating only a small subset of experts per token, and typically rely on a learned router to map hidden states to expert assignments.
arXiv:2603. 20895v3 Announce Type: replace-cross Abstract: Existing routers rely on semantic query features or handcrafted features, which often fail to capture model-specific failures or intrinsic task difficulty.
arXiv:2603. 24787v2 Announce Type: replace Abstract: Routing has emerged as a promising strategy for balancing performance and cost in large language model (LLM) systems that combine lightweight models with powerful but expensive large models.
arXiv:2609.01027v1 Announce Type: new Abstract: Out-of-distribution (OOD) detection predicts whether a test image belongs to none of the predefined classes. To evaluate this task, benchmarks need ima...
The paper introduces Ban&Pick, a post‑training, plug‑and‑play routing strategy for Sparse Mixture‑of‑Experts large language models. It identifies and reinforces a small group of highly influential experts while dynamically pruning redundant ones, leading to accuracy gains across math, code, and reasoning benchmarks. Experiments on DeepSeek and Qwen3 show notable performance improvements and a 1.25× inference speedup without retraining or architectural changes.
arXiv:2606. 06467v1 Announce Type: cross Abstract: Long-context inference in modern LLMs is increasingly constrained by decoding efficiency, especially in reasoning-heavy settings where models generate long intermediate chains of thought.
The paper presents a rigor‑matched audit comparing two periodic‑step, search‑based layer‑skipping methods for efficient large language model inference: a confidence‑gated early‑exit baseline (ConfLayers) and a self‑speculative decoding approach (SWIFT). Across two Qwen2.5 model scales and tasks (GSM8K reasoning and CNN/DailyMail summarization), SWIFT consistently outperforms ConfLayers in accuracy and, after separating search overhead, achieves faster true inference speed in most settings. The study also evaluates two trained‑routing methods (LayerRoute and LayerDrop), finding modest speedups but significantly lower accuracy, especially for LayerRoute on GSM8K at 1.5B.
arXiv:2610.16152v1 Announce Type: cross Abstract: Although post-training significantly enhances model performance, the use of massive undisclosed data raises concerns about benchmark contamination an...