AI Native Foundation

WorldReasonBench: Human-Aligned Stress Testing of Video Generators as Future World-State Predictors

2026-05-12

Geometry Conflict: Explaining and Controlling Forgetting in LLM Continual Post-Training

2026-05-12

Model Merging Scaling Laws in Large Language Models

2026-05-12

Auto-Rubric as Reward: From Implicit Preferences to Explicit Multimodal Generative Criteria

2026-05-12

Memory-Efficient Looped Transformer: Decoupling Compute from Memory in Looped Language Models

2026-05-12

X-OmniClaw Technical Report: A Unified Mobile Agent for Multimodal Understanding and Interaction

2026-05-12

Pixal3D: Pixel-Aligned 3D Generation from Images

2026-05-12

Key-Value Means

2026-05-12

G-Zero: Self-Play for Open-Ended Generation from Zero Data

2026-05-12

Dynamic Skill Lifecycle Management for Agentic Reinforcement Learning

2026-05-12

Rebellious Student: Reversing Teacher Signals for Reasoning Exploration with Self-Distilled RLVR

2026-05-12

LLaVA-UHD v4: What Makes Efficient Visual Encoding in MLLMs?

2026-05-12

Make Each Token Count: Towards Improving Long-Context Performance with KV Cache Eviction

2026-05-12

RigidFormer: Learning Rigid Dynamics using Transformers

2026-05-12

Prompt-Activation Duality: Improving Activation Steering via Attention-Level Interventions

2026-05-12

SlimSpec: Low-Rank Draft LM-Head for Accelerated Speculative Decoding

2026-05-12

SlimQwen: Exploring the Pruning and Distillation in Large MoE Model Pre-training

2026-05-12

Mela: Test-Time Memory Consolidation based on Transformation Hypothesis

2026-05-12

Addressing Performance Saturation for LLM RL via Precise Entropy Curve Control

2026-05-12

Shaping Schema via Language Representation as the Next Frontier for LLM Intelligence Expanding

2026-05-12

HF Papers

WorldReasonBench: Human-Aligned Stress Testing of Video Generators as Future World-State Predictors

Geometry Conflict: Explaining and Controlling Forgetting in LLM Continual Post-Training

Model Merging Scaling Laws in Large Language Models

Auto-Rubric as Reward: From Implicit Preferences to Explicit Multimodal Generative Criteria

Memory-Efficient Looped Transformer: Decoupling Compute from Memory in Looped Language Models

X-OmniClaw Technical Report: A Unified Mobile Agent for Multimodal Understanding and Interaction

Pixal3D: Pixel-Aligned 3D Generation from Images

Key-Value Means

G-Zero: Self-Play for Open-Ended Generation from Zero Data

Dynamic Skill Lifecycle Management for Agentic Reinforcement Learning

Rebellious Student: Reversing Teacher Signals for Reasoning Exploration with Self-Distilled RLVR

LLaVA-UHD v4: What Makes Efficient Visual Encoding in MLLMs?

Make Each Token Count: Towards Improving Long-Context Performance with KV Cache Eviction

RigidFormer: Learning Rigid Dynamics using Transformers

Prompt-Activation Duality: Improving Activation Steering via Attention-Level Interventions

SlimSpec: Low-Rank Draft LM-Head for Accelerated Speculative Decoding

SlimQwen: Exploring the Pruning and Distillation in Large MoE Model Pre-training

Mela: Test-Time Memory Consolidation based on Transformation Hypothesis

Addressing Performance Saturation for LLM RL via Precise Entropy Curve Control

Shaping Schema via Language Representation as the Next Frontier for LLM Intelligence Expanding

About

Insights

Case Study

Legal