Found 42,749 document(s)
Page 19 of 2850
-
As LLM adoption surges, scaling model serving requires unlocking stranded GPU capacity alongside deploying more GPUs. Speculative decoding brings two models into the serving process: a small draft model proposes tokens t...
-
Large language models sometimes reply in an unexpected language, a failure known as language confusion. We ablate individual attention heads in 12 model checkpoints spanning seven model series, using FLORES-200 continuat...
-
Fine-tuning language models on consumer GPUs is constrained by computation and memory. Although LoRA reduces trainable parameters and optimizer state, frozen projections remain expensive, and activations retained for bac...
-
We introduce a novel triple-overlap implementation of the pipelined preconditioned conjugate gradient (Pipe-PCG) and pipelined stabilized biconjugate gradient (Pipe-BiCGStab) solvers, methods widely used in scientific ap...
-
Flow Matching (FM) is an efficient simulation-free approach for training flow-based generative models, making it a promising foundation for scalable time-series anomaly detection (TSAD). However, standard FM learns a mon...
-
Policy Is Not Capability: Some Safety Benchmarks Measure Which Definition of "Unsafe" You Trained OnA guardrail classifier is reported as if unsafe were a property of text. It is not: it is a property of a policy, and the policy enters the model through its labels. We exploit an accident of an industrial deployment in ...
-
Generative recommendation constructs semantic IDs for catalog items and predicts the next ID from user interaction histories. However, existing methods face limitations in both stages. First, conventional quantizers may ...
-
An agent can retain information before receiving a future question and later retrieve evidence from an archive. Evaluating this arrangement requires accounting for both retained information and the width of each retrieva...
-
Repository-level retrieval-augmented generation systems must keep vector indexes synchronized with continuously evolving code, yet there is no principled basis for deciding which changed code units actually need new embe...
-
MODIS surface reflectance (SR) products often suffer from data gaps due to cloud contamination and sensor malfunctions, limiting their applications in long-term and large-scale studies. Most existing methods rely on clou...
-
Retrieval-augmented generation (RAG) is the dominant paradigm for grounding large language models in external corpora, but its similarity-driven, chunked pipeline has structural blind spots: it cannot anticipate which pa...
-
Cryptographic API misuse in large language model-generated code can compromise software security. While many existing approaches improve code security by modifying model parameters or inputs, how representation changes a...
-
Advancing physics reasoning in large language models requires scalable training data that connect physical laws, concrete problem scenarios, and verifiable numerical answers. Expert-designed problems offer top-down contr...
-
Traffic forecasting models based on spectral graph convolution offer an interpretable account of how information is mixed between sensors, but they rely on an eigendecomposition of the graph Laplacian. It costs cubic tim...
-
Catastrophic forgetting in sequential deep learning arises from a fixed-capacity architectural constraint: parameter updates encoding new tasks perturb weight subspaces supporting prior representations. We introduce the ...