Trending
What is gaining attention fastest, ranked by growth in upvotes as reported by the source, plus brand-new entries that arrived with traction. Looking for everything that just appeared, in order? See New.
- Waypoint-OPD: Transferring RL-Induced Policy Shifts via Gated On-Policy Distillation New+123 upvotes
Policy-shift distillation reuses the difference between a teacher's log probabilities before and after reinforcement learning (RL) as supervision on student-gen
- Speculative Forcing for Efficient Autoregressive Video Generation New+58 upvotes
Few-step autoregressive video generators enable streaming generation, yet they run every token through the final denoising step, regardless of how much refineme
- SCOPE: Semantic Bounded Credit Optimization for Prompt Engineering in Text-to-image Generation New+44 upvotes
Reinforcement learning (RL) enables language models to improve text-to-image (T2I) generation by rewriting prompts for a frozen image generator. Existing method
- 3DCodeVerse: Exploring 3D Worlds as Code New+41 upvotes
We introduce 3DCodeVerse, a research ecosystem that enables coding agents to create editable and verifiable 3D content at scale. 3DCodeVerse comprises large-sca
- Mosaic: Multi-Session Agentic Reinforcement Learning New+26 upvotes
Most existing agentic reinforcement learning methods assume single-session trajectories, where interaction proceeds within one continuously accumulated context.
- AutoSFT: An Agentic System for Automated Data Research in Supervised Fine-Tuning New+22 upvotes
Language model agents increasingly automate post-training, but when given the full training stack they converge on supervised fine-tuning and spend most effort
- MLExpert: Towards a Generalist Agentic Model for Autonomous Machine Learning Engineering New+21 upvotes
Large language models (LLMs) are expanding from assisting scientific research and engineering to autonomously developing and improving AI systems, advancing AI
- Routing Policy Optimization for Online Diffusion Reinforcement Learning New+18 upvotes
Reinforcement learning can align visual generators with reward signals beyond their pretraining objectives, but existing score-function methods typically introd
- Intern-Decision: Efficient Multimodal Decision Making with Autoregressive Masked Language Modeling New+16 upvotes
Interactive agents repeatedly choose among a finite set of actions, labels, or ordinal outcomes. Generating a free-form response for each choice adds latency an
- Terminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-Refine New+16 upvotes
Converting verifier outcomes into useful policy updates and partial progress into complete task resolution is crucial for terminal agents. We present Terminal-R
- JevAgentHarness: Benchmarking, Deploying and Training Decision Models That Share Control with LLMs New+14 upvotes
Large language model (LLM) agents now fix software, operate computers and serve customers, but their harnesses, the programs that run them, call a large LLM at
- Velox: Spatial Memory Makes World Models Efficient New+13 upvotes
Video world models turn actions and camera motion into long visual rollouts, making video a natural interface for interactive simulation. Such interaction requi
- Each Reward to Its Relevant Tokens: Reward-Specific Token-Level Credit Assignment for Multi-Reward RL New+13 upvotes
Language models are increasingly trained against several verifiable rewards at once, such as the correctness and the format of a tool call, or the individual co
- OpenGCUA: Toward a Fully Open Stack for Generalist Computer Use New+7 upvotes
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
- Learning Self-Orchestration in Decentralized Agent Swarms New+6 upvotes
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 fix
- Memory as Weights: Internalizing Long-Term History for Streaming Videos New+6 upvotes
Streaming vision-language models (VLMs) must continuously integrate information from the past while remaining responsive to current observations and user querie
- CarryOn: A Dual-Loop Harness for Personalized and Reliable Robot Assistance New+6 upvotes
While generalist vision-language-action (VLA) models have enabled diverse open-world manipulation, real-world robotic assistance requires two critical capabilit
- MJev-Spatial: System One Spatial Intelligence via Native Attention Matching New+3 upvotes
Spatial intelligence lets an agent perceive and act in the physical world, and many of its judgments must be made at interaction rates: which spatial relation h
- CoDA: Co-Evolution of Data Engines and Action Policies New+2 upvotes
Generative simulation can automate robot environment construction, but physically valid environments are not necessarily useful for training. The challenge is t
- MMSearch-OPD: Easy Made Hard, Hard Made Learnable for Long-Horizon Multimodal Search Agents New+2 upvotes
Multimodal deep search is moving from short queries to long-horizon investigations that continually acquire evidence. However, training such agents remains cons
- Trajectory-Level Speculation with Semantic Verification for Tool-Using Agents New+1 upvotes
Large language model agents solve tasks through long sequences of tool calls, and the prevailing design has a single strong model generate every step, although
- A team of Gemini 4 Argon agents beats Minecraft New+1 upvotes
We placed four Gemini 4 Argon agents in a sandboxed Minecraft survival world and gave them one goal: defeat the Ender Dragon. Each agent controlled a player via
- EVA-Cap: Optimizing Audiovisual Video Captioning via Event-Centric Alignment New+1 upvotes
Recent advances in Omni Language Models (OLMs) provide a unified solution for audiovisual video captioning. Nevertheless, existing approaches predominantly trea
- Skill-Decoupled On-Policy Expert Distillation for Unified Vision-Language Models New+1 upvotes
Reinforcement learning (RL) can elicit specialized visual capabilities in vision-language models (VLMs), yet jointly optimizing heterogeneous objectives often i
Just indexed
- Connecting…
Counts are the numbers reported by Hugging Face, GitHub and arXiv-linked communities, recorded daily by Silytics. Downloads on Hugging Face are 30-day figures, so a model that is already popular can still rank. Papers are ranked by Hugging Face daily-paper upvotes.