The Geometry of Reasoning: Flowing Logics in Representation Space
arXiv:2510.09782v1 Announce Type: new Abstract: We study how large language models (LLMs) “think” through their representation space. We propose a…
SPG: Sandwiched Policy Gradient for Masked Diffusion Language Models
arXiv:2510.09541v2 Announce Type: replace-cross Abstract: Diffusion large language models (dLLMs) are emerging as an efficient alternative to autoregressive models due…
video-SALMONN S: Streaming Audio-Visual LLMs Beyond Length Limits via Memory
arXiv:2510.11129v1 Announce Type: cross Abstract: Continuous, high-frame-rate, high-resolution processing of long video streams is critical for future AI agents, yet…
PhysioME: A Robust Multimodal Self-Supervised Framework for Physiological Signals with Missing Modalities
arXiv:2510.11110v1 Announce Type: cross Abstract: Missing or corrupted modalities are common in physiological signal-based medical applications owing to hardware constraints…
ChoirRec: Semantic User Grouping via LLMs for Conversion Rate Prediction of Low-Activity Users
arXiv:2510.09393v2 Announce Type: replace-cross Abstract: Accurately predicting conversion rates (CVR) for low-activity users remains a fundamental challenge in large-scale e-commerce…
Hypothesis Hunting with Evolving Networks of Autonomous Scientific Agents
arXiv:2510.08619v1 Announce Type: new Abstract: Large-scale scientific datasets — spanning health biobanks, cell atlases, Earth reanalyses, and more — create…
On the optimization dynamics of RLVR: Gradient gap and step size thresholds
arXiv:2510.08539v2 Announce Type: replace-cross Abstract: Reinforcement Learning with Verifiable Rewards (RLVR), which uses simple binary feedback to post-train large language…
CapGeo: A Caption-Assisted Approach to Geometric Reasoning
arXiv:2510.09302v1 Announce Type: cross Abstract: Geometric reasoning remains a core challenge for Multimodal Large Language Models (MLLMs). Even the most…
CLARity: Reasoning Consistency Alone Can Teach Reinforced Experts
arXiv:2510.09278v1 Announce Type: cross Abstract: Training expert LLMs in domains with scarce data is difficult, often relying on multiple-choice questions…
Learning Neural Exposure Fields for View Synthesis
arXiv:2510.08279v2 Announce Type: replace-cross Abstract: Recent advances in neural scene representations have led to unprecedented quality in 3D reconstruction and…