Signal over volume.

LongTraceRL: Enhancing Long-Context Reasoning via Trajectory-Based Distractor...

LongTraceRL: Enhancing Long-Context Reasoning via Trajectory-Based Distractors and Rubric Rewards

BaLoRA: Accelerating LoRA Convergence via Balanced Parameter Manifolds

Stateful Online Monitoring: Defending Against Distributed Agentic Cyberattacks

Graphical Einops: A Formal Calculus for Tensor Programming and Attention Opti...

AsyncTool: A New Benchmark for Evaluating LLM Agents in Asynchronous Multi-Ta...

How LoRA Remembers: Uncovering the Parametric Memory Law and Phase Transition...

LaRA: Detecting Data Contamination in RL Post-Training via Layer-wise Represe...

LiteCoder-Terminal: Scaling Long-Horizon Terminal Agents via Zero-Dependency ...

DarkForest: Reducing Multi-Agent LLM Communication for Higher Accuracy and Lo...

Collaborative Parallel Thinking: Reducing Redundancy in LLM Test-Time Scaling

IC-Seg: Resolving Ambiguity in Referring Segmentation via Multi-turn Clarific...

Your Agents Are Aging Too: New Benchmark Reveals Reliability Decay in Long-Li...

Current Video Quality Models Fail to Accurately Assess Diffusion-Based Super-...

View Dropout: Making Visual Thinking Load-Bearing in Multimodal Models

GUI-CIDER: Mid-training GUI Agents via Causal Internalization and Density-awa...

GE-Sim 2.0: Closing the Loop with a High-Fidelity Video World Simulator for R...

MM-CreativityBench: Testing and Improving Creative Tool Repurposing in Multim...

STREAM: Mining High-Value Task-Oriented Dialogues from Streaming Media

Gemini Embedding 2: A Unified Native Multimodal Embedding Model

Modeling Agentic Technical Debt and Stochastic Tax: A New Framework for AI Go...

Claw-Eval-Live: A Live Benchmark for Evolving Real-World Workflow Agents

Crab: Semantics-Aware Checkpoint/Restore Runtime for Agent Sandboxes

Skills-Coach: Automated Self-Evolving Skill Optimizer for LLM Agents via Trai...

ML maps Vicsek model phase diagram: K-Means clustering + NN classifier reveal...