Papers
Videos
Whiteboards
Open Problems
Email Digest
Labs
Pricing
Updates
Log in
Sign up
Research Explorer
Papers
Videos
Whiteboards
Open Problems
Email Digest
Labs
Pixel Art Bench
Self-Improving Tweets
AI Pixel Art Generator
All Projects
Pricing
Log in
Sign up
Discord
Discord Logo Streamline Icon: https://streamlinehq.com
Whiteboard Explanations
Scientific Theory of Deep Learning
Decoupled DiLoCo: Resilient Distributed Pre-training
Hyperloop Transformers: Efficient Language Modeling
Soft-Label Governance in Multi-Agent Systems
Adversarial Multi-Agent LLM Defect Review
Scaling Self-Play with Self-Guidance
Convergent Evolution: LM Number Representations
Generalization at the Edge of Stability
Agentic Forecasting with Bayesian Linguistic Beliefs
AI Agents Lack Scientific Reasoning
Cluster-Aware Upcycling for MoE Specialization
Neural Garbage Collection: Efficient LM Memory Management
VLA Foundry: Unified Vision-Language-Action Training
Scaling Test-Time Compute for Agentic Coding
QuantumQA: Physics-Consistent Scientific Reasoning
ConforNets: Latent Conformational Control in OpenFold3
LLMs Corrupt Documents in Delegation
Neural Gabor Splatting for High-frequency Reconstruction
Digitally Controlled Silicon QPU
The LLM Fallacy in Cognitive Workflows
BankerToolBench: AI in Investment Banking
KVCache in Cross-DC LLM Serving
Claude Code: Design of AI Agent Systems
TokenGS: 3D Gaussian Prediction with Tokens
GlobalSplat: Efficient 3D Gaussian Splatting
Value Gradient Flow in Reinforcement Learning
Memory Transfer Learning in Coding Agents
LongCoT: Benchmarking Long-Horizon Reasoning
MAGICIAN: Long-Term Planning for Active Mapping
AiScientist: Autonomous ML Research
Humanoid Manipulation with Touch Dreaming
Lyra 2.0: Generative 3D World Exploration
Nucleus-Image: Sparse MoE for Image Generation
TIPSv2: Enhanced Patch-Text Alignment
Multi-User Large Language Model Agents
SD-Zero: Dense Supervision from Binary Rewards
Parcae: Scaling Laws for Stable Looped LMs
Transformers Plan with Multi-Token Prediction
Mechanistic Dynamics of Looped Transformers
Zero-shot World Models: Efficient Visual Learners
Dimension Descent for PMT in High Dimensions
Evaluating Thought Streams in Gemini VLMs
Adaptive Subword Segmentation Dynamics
Physics Olympiad via RL on Simulators
ExecTune: Steering Black-Box LLMs with Guide Models
3D Transport Chemistry on K2-18b
Cross-lingual Transfer in Low-Resource African Languages
LLMs: Unified Mechanism for Harmful Generation
Efficient RL Training via Experience Replay
Elastic Looped Transformers for Visual Generation
Nanohertz GW and Cosmic Structure Equilibrium
Sparse Diffusion for Open-World Motion Forecasting
IatroBench: Omission Harm in AI Safety
Counterfactual Density Effects and the German East--West Income Gap
The Economics of War: Militarization and Growth in an AK Economy
Forecasting and Manipulating the Forecasts of Others
LLM Training as Lossy Compression
Depth Ceiling: Limits of LLM Latent Planning
Malicious Intermediary Attacks on LLM Supply Chain
ClawBench: AI Agents on Real-World Web Tasks
PIArena: Platform for Prompt Injection Evaluation
Exponential Quantum Advantage in Data Processing
Scal3R: Scalable 3D Reconstruction
Neural Computers: Unified Learned Runtime
MARS: Multi-Token Generation in AR LMs
Target Policy Optimization
PaperOrchestra: Multi-Agent AI Paper Writing
AI and the Structure of Mathematics
Benchmarking Agentic Skills in LLMs
ClawsBench: LLM Agents in Simulated Workspaces
In-Place Test-Time Training for LLMs
Boxer: Robust 2D-to-3D Lifting for Open-World Objects
Gym-Anything: Automated Agent Environments
MegaTrain: Single-GPU Training of 100B+ LLMs
AvatarPointillist: 4D Gaussian Avatarization
FlashSAC: Fast, Stable Off-Policy RL
Free-Range Gaussians: Non-Grid 3D Reconstruction
AI Assistance Harms Persistence & Performance
Solving Unknown-Difficulty Problems
Delta Tokens: Efficient Generative World Modeling
LoMa: Local Feature Matching Revisited
AI Agents as Corporate Accomplices
GameSight: Knowledge-Enhanced Soccer Commentary
Security Evaluation: OpenClaw and Its Variants
Open-Source LiDAR & Monocular Navigation
SenseMath: Evaluating LLM Number Sense
TabEBM: Class-Specific Augmentation for Tabular Data
Test-Time Scaling: Overtraining for Optimal Compute
BraiNCA: Neural Cellular Automata for Morphogenesis
OmniSimpleMem: Lifelong Multimodal Agent Memory
Self-Preservation Bias in LLMs
ASI-Evolve: Agentic AI Research Framework
Rising Tides in AI Labor Automation
Simple Self-Distillation for Code Generation
ByteRover: Agent-Native Memory for LLMs
Skill0: Agentic RL for Skill Internalization
Multimodal Analysis of Israel-Hamas War on YouTube Shorts
Screening Is Enough: Multiscreen Model
Nuclear Test-Transient Link in POSS-I Plates
Transformers and Attention for Applied Mathematics
Show More
