OpenGCUA: Toward a Fully Open Stack for Generalist Computer Use
Computer use has reached a turning point: the focus is expanding from GUI-only interaction to systems that complete tasks on users' computers through visual and programmatic interf...
Papers
CoDA: Co-Evolution of Data Engines and Action Policies
Generative simulation can automate robot environment construction, but physically valid environments are not necessarily useful for training. The challenge is to identify environme...
Kolab: Benchmarking Human-Agent Collaboration in Long-Horizon Workflows
When humans collaborate with AI agents on long-horizon tasks, what happens to the agents' contributions? For solving long-horizon tasks, humans place greater importance on ongoing ...
Stuck on Shuffle: On Attribute-Object Binding in Vision-Language Models
Retrieval is a fundamental task in multimodal systems: given an image or text query, the goal is to find the most relevant images or texts in a collection, often using a shared emb...
PremiseCheck: Efficient Fact-Checking of Human and AI Peer Reviews
Peer review is under strain from the growth of submissions, and review comments written by humans or AI can misstate what a paper says. We introduce ReviewFact, a benchmark that la...
Can Visual Generation Expand What Unified Multimodal Models Know?
Unified multimodal models (UMMs) receive image-generation supervision that exposes them to visual information not fully documented in text. Can learning to generate images expand w...
Skip the Compressor: Coding Agents Fix More Issues When They Pick Whole Files
A coding agent works on repositories that can be far larger than its context window can hold. Prompt compressors and retrievers offer solutions to this problem, but patching an iss...
AVC-Reward: Event-Centric Audio-Video Consistency Evaluation
As joint audio-video generation models become more capable, they can generate audio-visual scenes exhibiting three forms of complexity: multi-event composition, multi-layer audio m...
Behavioral Compositionality of Sparse Autoencoder Representations
Sparse autoencoders (SAEs) represent model activations with sparse, linearly decoded features, but the downstream effects of those features need not compose additively. We introduc...
Resisting Harmful Instructions through Training-Free Weight Editing
Instruction-following language models can comply with requests that conflict with intended behaviors and safety alignment. We introduce a training-free weight-editing approach for ...
Mine, Sample, Render: Test-Time Diversity Alignment Of Generative Models
Modern generative models produce high-quality individual outputs but struggle to sample diversely, even when explicitly instructed to do so. We introduce MINE-SAMPLE-RENDER (MSR), ...
Selection Is Not Learning: Choosing Model Futures
Self-evolving language models train on tasks they generate, so every round must decide which updated model the next round inherits. Current loops such as R-Zero, Agent0, and Skill ...
OmniXRef: A Unified Framework for Multimodal Multi-Reference Video Editing
Creative intent in video editing can be expressed through text instructions (e.g., "remove the people"), image references (e.g., subject appearance), video references (e.g., camera...
Beyond Compilation: Evaluating Faithful Natural-Language-to-Lean Statement Formalization
Lean verifies that a generated declaration is well typed, but cannot check whether it states what the user intended. We measure how far compilation overstates faithfulness in state...
LEASH: Uncertainty-Aware Weak-to-Strong Escalation and Oversight for Computer-Use Agents
Computer-use agents increasingly tackle long-horizon tasks, where running a frontier model at every step is expensive and a small open-weight model alone often fails. Weak-to-stron...
RealCLI: Toward a Unified Evaluation of Human-Agent Collaboration for CLI Agents
Command-line interface (CLI) agents increasingly execute longer-horizon tasks, yet existing benchmarks either assume that user instructions remain fixed throughout execution or eva...
When Morphs Look Better: Face Image Quality as a Training-Free Signal for Morphing Attack Detection
Face image quality assessment (FIQA) estimates the utility of a face image for automated face recognition and has been proposed as a morphing attack detection (MAD) signal that req...
Representation Redistribution in Sparse Autoencoders
Sparse autoencoders (SAEs) decompose language model activations into sparse features, but prompt transformations designed to preserve the underlying request can redistribute activi...
First Passage Attention
Augmenting softmax attention with forget gates typically relies on summing decay contributions along the key--query path. Because such accumulation is monotonic, these mechanisms c...
Learning Self-Orchestration in Decentralized Agent Swarms
The capabilities of an agent swarm depend on both its individual members and how they coordinate. Yet existing multi-agent systems often train agents within fixed orchestration str...
TokenEconomizer: Experience-Driven Black-Box Cost Control for Coding Agents
Advances in frontier coding agents have sharply increased the token costs of AI-assisted software engineering. Their token efficiency is typically evaluated on isolated tasks, wher...
Mosaic: Multi-Session Agentic Reinforcement Learning
Most existing agentic reinforcement learning methods assume single-session trajectories, where interaction proceeds within one continuously accumulated context. However, agentic te...
DialFork: Going Beyond Factuality to Evaluate Procedural Hallucination in Conversational Agents
Large language models (LLMs) deployed as conversational agents are increasingly expected to follow procedural logic. While factual hallucination has received extensive attention, a...
Gen4P: Video Generators are 4D Perceivers
We introduce Gen4P, a unified framework for recovering 4D geometry and motion from monocular video. We show that a pretrained video generator can be repurposed into a feed-forward ...