Signal over volume.

K-BROWSECOMP: A New Benchmark for Korean-Context Web Browsing Agents

Beyond Correctness: Auditing Source-Dependence in Multi-Source Medical RAG Sy...

Beyond Correctness: Auditing Source-Dependence in Multi-Source Medical RAG Systems

τ0-WM: A Unified Video-Action World Model for Robust Robotic Manipulation

Linear Ensembles Wash Away Watermarks: The Fundamental Fragility of AI Attrib...

Beyond Text: Introducing VISUAL SKILL for Multimodal Agent Capability

SOCO: A New Taxonomy-Driven Benchmark for Semantic Object Correspondence

AI, Take the Wheel: Uncovering the Drivers of Trust and Delegation in Human-A...

Harness-1: Offloading Search Bookkeeping to Environment State Boosts Retrieva...

LVSA: Accelerating Long Video Diffusion with Training-Free Sparse Attention

The Hamilton–Jacobi Theory of Deep Learning: A Unifying Mathematical Framework

MineExplorer: A New Benchmark for Evaluating MLLM Agents in Open-World Minecr...

Who Annotates in NLP? A Large-Scale Audit of Human Annotation Reporting (2018...

PARCEL: Solving the Spectral Bottleneck in Elastic Vision-Language Models

3DCodeBench: A New Standard for Evaluating VLM Agents in Procedural 3D Modeling

Temporal Scheduling: A New Dimension for Optimizing LLM Reasoning via RLVR

OpenWebRL: Scaling Visual Web Agents via Online Multi-turn Reinforcement Lear...

LongLive-RAG: Enhancing Long Video Consistency via Latent Retrieval

ChartArena: A Unified Bilingual Benchmark for Robust Chart and Diagram Parsing

TASTE: Reversing Task Construction to Build Harder, More Diverse Agent Benchm...

EVA01: Unifying 3D Understanding, Generation, and Multi-Turn Editing in a Sin...

Physical Admissibility Gate: Ensuring Robot AI Predictions Respect Real-World...

FINEVERIFY: Scaling Agentic Search via Fine-Grained Sub-Question Verification

PaSBench-Video: Exposing the Failure of MLLMs in Proactive Safety Warning

HLL Benchmark: Testing if Multimodal Agents Can Cross Humanity's Last Line of...