VASPO: Self-Distillation with Privileged Value Estimation
On-policy self-distillation provides dense supervision for reasoning models via a privileged self-teacher. However, this supervision reflects local token preferences rather than th...
Research Papers
Latest AI/ML research from ArXiv and Papers with Code
Robust Explanatory Analysis for Structured Optimization Networks (REASON) in Learning
We seek a hierarchical resolution of residual error under a composite learning objective as a principled mechanism for identifying and exploiting additional information, particular...
Diminishing Returns of Data in Sparse Autoencoders: Functional Behavior does not Guarantee Feature Reproducibility
Sparse autoencoders (SAEs) are typically trained on hundreds of millions to billions of activation tokens, yet little work has tested whether this scale is necessary. We study how ...
WorldCast: Exploring Intelligent Visual Generation in World State Prediction
Humans interpret multimodal observations of the world and forecast how it will evolve. Recent multimodal generative systems have demonstrated similar capabilities: they can integra...
One Bias Score, Two Factors: How Mitigations Affect Option Bias in Multiple-Choice Evaluation
Reordering multiple-choice options can change a model's answer, an effect studied as selection bias (a preference for certain option letters) and as order sensitivity (changes in a...
BizBench: Unveiling Hidden Challenges in Deploying Long-Horizon Agents for Real-World Business
Existing long-horizon agent benchmarks often emphasize final task success while providing limited assessment of intermediate execution validity. Yet real-world business operations ...
Orbia: Do Generated Videos Form a Consistent 3D World?
Video generators serve as interactive worlds for exploration and simulation, yet realistic and visually appealing frames do not necessarily depict a consistent 3D world. Existing b...
What Generation Knows: Unlocking Fine-Grained Understanding in Unified Multimodal Models
Fine-grained visual understanding requires distinguishing subtle differences in local attributes and spatial relations, yet remains challenging for existing vision-language models....
Intern Lumina U2: A Multi-Codebook Diffusion Large Language Model for Omni-Visual Understanding and Image Generation
Unified multimodal generation covers discrete, continuous, and hybrid paradigms. Discrete approaches with single-codebook visual tokenizers align naturally with text modeling but h...
MAI-Gym: An Open All-in-One Platform for GUI Agent Evaluation and RL Training across Mobile and Computer Environments
Online reinforcement learning (RL) with verifiable rewards has driven rapid progress in mathematics, coding, and tool use, but scaling this success to GUI agents remains challengin...
SERO: Learning from Executable Oracle Continuations for Long-Horizon Agentic Reinforcement Learning
Long-horizon agentic RL is constrained by the futures visited by the current policy. When rollouts fail to reach useful behaviors, improved credit assignment cannot provide supervi...
VITAL: From Visual Motion to Precise Action
World-action models aim to turn visual anticipation into driving decisions, yet typically train action experts on trajectory representations defined independently of visual dynamic...
SafeDriveVLA: Navigation-Conditioned World Model Dreaming for Conflict-Aware End-to-End Autonomous Driving
Vision-Language-Action (VLA) models are gaining attention in autonomous driving, inheriting commonsense priors from large language models (LLMs) and widening the navigation interfa...
Equal CER, Unequal Damage: Severity-Weighted Character Error Rate for OCR
Character Error Rate (CER), the standard metric for OCR and vision-language models (VLMs), charges a typo as much as a wrong name. We introduce semCER, which keeps CER's alignment ...
ActionOPD: Distilling Decisions, Not Tokens
Distilling capable language-model agents into smaller models requires aligning the student's behavior with the teacher's. This raises a fundamental question: at which states and at...
Learning to Read Out: Unembedding Dynamics in Language Model Pretraining
When does a language model acquire a capability? Developmental studies usually answer this by tracking when models start producing the correct token. But that prediction passes thr...
Self-Conditioning Enables Deep Sequential Reasoning in Flow Matching
Recent work has explored visual reasoning through image generation, where models solve visual-reasoning problems by directly generating solution images. Intuitively, iterative samp...
VLX-Go: Robust Embodied Tracking through Multi-Stage Learning and On-Policy Refinement
Embodied visual tracking requires a robot to identify a designated target from visual observations and control its own motion to follow that target through an environment. Followin...
Recovering Reliable Supervision via Critique-Guided On-Policy Distillation
On-policy distillation (OPD) trains a student model on its own rollouts with token-level supervision from a teacher, assuming that the teacher can reach a correct answer from any p...
MolLean: A Process-Verifiable Language for Chemical Reasoning
Chemical chain-of-thought can contain hallucinated structural and property claims. We thereby introduce MolLean, a process-verifiable language that grounds chemical reasoning in an...
The COUNT: Counting N-grams Unlocks Scaling on Host Memory and Training-Free Adaptation
The scaling relationship of compute and performance for language models (LMs) has been extensively studied; more recently, models are being designed to scale memory use, as well. I...
Any Time, Any Space: A World-State-Centric Framework for Agentic Video Generation
Long-form agentic video generation must turn a screenplay into shots that remain spatially and temporally consistent across nonconsecutive requests. Passing prompts or recent frame...
MorphAct: Robot Policies Morphing with the Worlds They Act In
Progress toward general intelligence in the physical world requires scaling robot learning across diverse environments and interactions. However, existing robot policies use shared...
LatentLoop: Scaling Looped MoE Models with Asymmetric Compute Allocation
Looped Mixture-of-Experts (MoE) models combine parameter reuse with sparse computation to improve performance under matched budgets. Prior work demonstrates that different loop vis...