Category
3,629 papers
This work proposes Drift-Aware Sparse Routing (DRS), a nonstationary sparse contextual routing with multiple knapsack constraints and an optional shadow-audit stream that evaluates a small fraction of prompts on several models.
Controlled fine-tuning, distillation, and agentic consistency-checking support the same conclusions, and the SciTrue team's participation in both subtasks of the NTCIR-19 SciClaimEval task is described.
3R-Bench (Refusal, Repetition, and Revision), a benchmark of 150 real-world cybersecurity requests augmented with two adversarial conversational settings, is introduced and eight LLMs are evaluated, finding that prior assistant behavior strongly changes responses to an unchanged request.
PARSER is a new residual sparsification method that shifts the compression objective from minimizing isolated matrix errors to preserving the expert output error, and achieves this by introducing output importance, which measures the actual contribution to the expert output error.
Reach audiences
Advertise in front of researchers, engineers, and readers.
The Evidence-Grounded Typed Knowledge Graph (EGT-KG), a retrieval framework to improve information retrieval with local SLMs is proposed, showing that EGT-KG outperforms the vanilla RAG method in most settings, with the best improvement from llama3:8b.
GLANCE is presented, the first one-pass block drafter that is lossless on an unmodified VLM target, and it breaks the cycle at both ends of a self-defeating cycle.
Requiring no training or auxiliary computation, stride-k subsampling exploits Whisper's preprocessing redundancy, indicating that its audio-token interface carries more capacity than downstream tasks require.
SIR is presented, a black box IPI attack that composes stealthy injections from a small library of reusable principles stated in plain language and wraps composition in an iterative feedback loop that diagnoses the victim's failed trajectories and distills the bypasses into new, named strategies that are reapplied acro...
CoVeil is proposed, a defense mechanism which dynamically optimizes transmitted signals to suppress leakage during decoding time while preserving the collaborative quality, and consistently improves the privacy-utility trade-off over existing baselines by reducing data leakage.
Despite its compact active-parameter budget, Turing-20B-A2B achieves, at the base-model stage, overall general capability exceeding Qwen3-8B Base and approaching Qwen3.5-9B Base, while maintaining strong long-context performance and favorable prefill-latency scaling.
Results show that medical knowledge updating depends not only on the update algorithm, but also on how knowledge is structured as supervision, and suggest that EMQ gives the most stable external transfer and retention among same-budget SFT variants.
LiteSearch-VL is a low-compute recipe for Qwen3-VL-2B and Qwen3-VL-4B that uses only released OpenSearch-VL trajectories, parameter-efficient LoRA adapters, and synthetic step-level preferences, identifying answer verification, not search depth, as the next bottleneck for small multimodal agents.
A Blog post by NVIDIA on Hugging Face
Data Mining & Modeling
A Blog post by Ai2 on Hugging Face
A Blog post by IBM Research on Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
A Blog post by IBM Granite on Hugging Face
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
