arXivarxivSilver Rate Is (Almost) Optimal for Gradient Descent Accelerationarxivmath.OCcs.LG6 days ago▲ 43Read story→
arXivarxivCopying explains the collective behavior of AI agents in the wildarxivcs.MAcond-mat.stat-mech6 days ago▲ 43Read story→
arXivarxivStudying Image Tokenizers as Visual Languages in Unified Multimodal Modelsarxivcs.CVcs.CL6 days ago▲ 43Read story→
arXivarxivNOAH: Learning the Full Patient Journey. A Longitudinal Multimodal Time-Aware Model for Representation and Forecastingarxivcs.LGcs.AI6 days ago▲ 43Read story→
arXivarxivA Data-Driven Framework for Identifying and Prioritizing RPA Opportunities in Healthcare Processesarxivcs.AIcs.CL6 days ago▲ 43Read story→
arXivarxivEntropy-Regularized Rank-Masked Policy Optimization for Test-Time Reinforcement Learning in Code Generationarxivcs.LGcs.CL6 days ago▲ 43Read story→
arXivarxivCo-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Failsarxivcs.AIpublisher:arxiv6 days ago▲ 43Read story→
arXivarxivExecCritic: Learn to Test, Test to Improve for Coding Agentsarxivcs.AIcs.CL6 days ago▲ 43Read story→
arXivarxivNearly Tight Rademacher Bounds for Sparsely Activated Neural Networksarxivcs.LGstat.ML6 days ago▲ 43Read story→
arXivarxivCanonical Color as a Lens into Concept Decodability in Vision Encoders and VLMsarxivcs.CVcs.AI6 days ago▲ 43Read story→
arXivarxivDeCAL: Towards Physically-Grounded Dexterous Vision-Language-Action Models via Contact-Aware Latent Co-Imaginationarxivcs.ROcs.AI6 days ago▲ 43Read story→
arXivarxivWhen Does Scale-Invariant Optimization Become Unstable? An Exact Schedule Law with Weight Decayarxivcs.LGpublisher:arxiv6 days ago▲ 43Read story→
arXivarxivMeClear: Cooperative Game-Theoretic Attribution and Risk-Aware Memory Clearance for Long-Horizon LLM Agentsarxivcs.AIcs.SE6 days ago▲ 43Read story→
arXivarxivSAEScientist-Bench: Can AI Agents Conduct Autonomous SAE Interpretability Research?arxivcs.AIcs.CL6 days ago▲ 43Read story→
arXivarxivCurriculum Learning as Transport: Understanding Curricula with Wasserstein Geodesicsarxivcs.LGpublisher:arxiv6 days ago▲ 43Read story→
arXivarxivThe Surprising Effectiveness of Approximate Value Iteration in Self-Playarxivcs.AIpublisher:arxiv6 days ago▲ 43Read story→
arXivarxivMeasuring LLM Sycophancy under Sustained Multi-Turn Pressurearxivcs.CLcs.AI6 days ago▲ 43Read story→
arXivarxivIt's Not RoPE that Creates Sinks: The Role of Self-Concentration and Value-Non-Mixing in Attentionarxivcs.CLpublisher:arxiv6 days ago▲ 43Read story→
arXivarxivGoDeep: Annotation-Free Open-Vocabulary 3D Scene Understanding via Language-Space Liftingarxivcs.CVcs.AI6 days ago▲ 43Read story→