Search: “routing”

arXiv:2605.07111v1 Announce Type: cross Abstract: Recent literature on fine-tuning Large Language Models highlights a fundamental debate. While Full Fine-Tuning (FFT) provides the representational pla

arXiv:2605.07846v1 Announce Type: new Abstract: Coarse-mask local image editing asks a model to modify a user-indicated region while preserving the surrounding scene. In practice, however, rough masks

arXiv:2602.03473v2 Announce Type: replace-cross Abstract: Continual learning, especially class-incremental learning (CIL), on the basis of a pre-trained model (PTM) has garnered substantial research i

arXiv:2605.01480v1 Announce Type: new Abstract: We study training-free image editing on Qwen-Image-Edit-2511, a 60-block multi-modal diffusion transformer (MMDiT) that concatenates noise and source-im

arXiv:2605.00572v1 Announce Type: new Abstract: Algorithm performance in combinatorial optimization is highly sensitive to parameter settings, while a single globally tuned configuration often fails t

arXiv:2604.23717v1 Announce Type: cross Abstract: Recent large audio language models (LALMs) demonstrate remarkable capabilities in processing extended multi-modal sequences, yet incur high inference

arXiv:2604.18866v1 Announce Type: new Abstract: Despite advances in object detection, aerial imagery remains a challenging domain, as models often fail to generalize across variations in spatial resol

arXiv:2604.17837v1 Announce Type: cross Abstract: An LLM's residual stream is both state and instruction: it encodes the current context and determines the next transformation. We introduce a paramete

arXiv:2604.16883v1 Announce Type: new Abstract: In long-context decoding for LLMs and LMMs, attention becomes increasingly memory-bound because each decoding step must load a large amount of KV-cache

arXiv:2604.09613v1 Announce Type: cross Abstract: Production vLLM fleets provision every instance for worst-case context length, wasting 4-8x concurrency on the 80-95% of requests that are short and s

arXiv:2604.09288v1 Announce Type: new Abstract: Trusted multi-view classification typically relies on a view-wise evidential fusion process: each view independently produces class evidence and uncerta

arXiv:2604.08075v1 Announce Type: new Abstract: Production vLLM fleets typically provision each instance for the worst-case context length, leading to substantial KV-cache over-allocation and under-ut

arXiv:2610.15595v1 Announce Type: new Abstract: In real-world indoor fire environments, visual degradation severely impairs instruction grounding, waypoint prediction, and long-horizon exploration in

arXiv:2610.17062v1 Announce Type: new Abstract: We propose a mitigation to reward hacking in reinforcement learning. Our setup involves a misspecified reward function that gives rise to unintended beh

arXiv:2610.18195v1 Announce Type: cross Abstract: For sensing-aided communications, single-endpoint sensing offers only a partial view of propagation, motivating link-centric sensing across both endpo

Training large language models (LLMs) is resource-intensive, and adapting them for diverse deployment scenarios with varying computational constraints remains challenging. While elastic architectures

arXiv:2610.04918v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards scales multimodal reasoning, but an outcome reward says how much a trajectory is worth, not how that valu

arXiv:2610.02270v1 Announce Type: cross Abstract: Medical vision-language model (VLM) evaluation is sensitive to workflow design, prompting strategy, and benchmark construction, yet most studies treat

arXiv:2610.03510v1 Announce Type: cross Abstract: Recent advances in autoregressive video generation have improved temporal consistency over extended durations, yet interactive storytelling requires m

arXiv:2610.01001v1 Announce Type: new Abstract: Model-based reinforcement learning (MBRL) can exploit simulated experience, but a simulator-to-target shift creates a model-selection problem: correctin

arXiv:2610.00912v1 Announce Type: new Abstract: Large language models can translate business descriptions into optimization models, but executable code may misrepresent constraints or objectives. A so

arXiv:2610.00309v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed in privacy-critical domains (e.g., healthcare, finance, and government), but their propensity t

arXiv:2610.01984v1 Announce Type: cross Abstract: Byte-level byte-pair encoding (BBPE) tokenizers are attractive for multilingual large language models (LLMs) because they cover all Unicode text. In U

arXiv:2609.37285v1 Announce Type: new Abstract: Laboratories often face a new molecular assay with 16-64 labels and a bank of predictors whose training data and parameters are unavailable. The practic

arXiv:2609.35831v1 Announce Type: new Abstract: Modern large language models now support context windows of more than one million tokens, which has raised the question of whether retrieval-augmented g

arXiv:2609.37831v1 Announce Type: new Abstract: Real-time diffusion-based video super-resolution (VSR) is in high demand for online streaming, yet stringent latency requirements often compromise gener

arXiv:2609.36926v1 Announce Type: cross Abstract: Recent power measurements provide valuable information for photovoltaic(PV) power forecasting, but directly extrapolating short-term trends can introd

arXiv:2609.33962v1 Announce Type: cross Abstract: Subject-independent motor-imagery (MI) EEG decoding can exhibit subject-level failures even when average performance appears acceptable: under subject

arXiv:2609.34722v1 Announce Type: new Abstract: Long-horizon camera-controlled video generation requires recovering previously observed content from an ever-growing visual history. Existing approaches

arXiv:2609.28475v1 Announce Type: new Abstract: Forecasting agents increasingly combine language-model reasoning, retrieval, ensembling, and calibration, but it remains unclear when each behavior shou

arXiv:2609.27373v1 Announce Type: new Abstract: Recurrent language models repeatedly apply shared network blocks to refine latent representations, but standard inference recomputes global attention at

arXiv:2609.26976v1 Announce Type: new Abstract: Learned context planning selects evidence atoms before an answer model reasons over them. We test whether this learned selection improves long-context m

arXiv:2609.22641v1 Announce Type: new Abstract: Autoregressive video world models enable temporally coherent generation for a single observer. Extending them to multiple agents requires consistency ac

arXiv:2609.22146v1 Announce Type: cross Abstract: We study how post-training changes the weights of Large Language Models (LLMs) relative to their pretrained weights. Across 12 post-training chains wi

arXiv:2609.24813v1 Announce Type: new Abstract: Vision-Language Models (VLMs) have demonstrated remarkable capabilities in multimodal tasks, yet they still exhibit poor ability in spatial reasoning. E

arXiv:2609.22214v1 Announce Type: new Abstract: Long multilingual conversational spoken question answering requires systems to balance long-range transcript semantics with sparse acoustic and speaker-

arXiv:2609.20299v1 Announce Type: new Abstract: Information aggregation across model layers are revealed to improve OOD detection. In contrast to crafting a method for layer-wise information aggregati

arXiv:2505.24846v3 Announce Type: replace Abstract: Reward modeling is a key step in building safe foundation models when applying reinforcement learning from human feedback (RLHF) to align Large Lang

arXiv:2609.14185v1 Announce Type: cross Abstract: Bird's Eye View (BEV)-based multi-view 3D object detection suffers from challenges of computational complexity, multi-scale feature extraction, and ef

arXiv:2609.14759v1 Announce Type: cross Abstract: Alignment applied after pretraining is shallow in a measurable way: a single direction in a model's residual stream can be edited out, and the model s

arXiv:2609.14614v1 Announce Type: new Abstract: Deep learning architectures remain rigidly built upon traditional fully connected layers. While networks scale up, few challenge this foundational root.

arXiv:2609.15160v1 Announce Type: new Abstract: Looped Transformers have recently demonstrated strong performance in both reasoning and language tasks by reusing a shared set of parameters across mult

arXiv:2609.12517v1 Announce Type: new Abstract: Long-Video Question Answering (LVQA) requires locating decisive evidence in hour-scale videos under a limited frame budget. Most training-free methods a

arXiv:2609.10346v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) process hundreds or thousands of visual tokens per image, incurring prohibitive inference costs. While existing

arXiv:2609.10372v1 Announce Type: new Abstract: We present the PACE, a framework for retrieval-augmented dialogue serving that formalizes Perceived Time-to-First-Response (PTFR) as a QoE objective and

arXiv:1706.04580v3 Announce Type: replace Abstract: The design and organization of complex robotic systems traditionally requires laborious trial-and-error processes to ensure both hardware and softwa