AI研究論文每日精選

MetaChain: A Fully-Automated and Zero-Code Framework for LLM Agents

History-Guided Video Diffusion

Steel-LLM:From Scratch to Open Source -- A Personal Journey in Building a Chinese-Centric LLM

APE: Faster and Longer Context-Augmented Generation via Adaptive Parallel Encoding

The Curse of Depth in Large Language Models

Lossless Acceleration of Large Language Models with Hierarchical Drafting based on Temporal Locality in Speculative Decoding

CustomVideoX: 3D Reference Attention Driven Dynamic Adaptation for Zero-Shot Customized Video Diffusion Transformers

Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scaling

Jakiro: Boosting Speculative Decoding with Decoupled Multi-Head via MoE

Towards Internet-Scale Training For Agents

Dual Caption Preference Optimization for Diffusion Models

LM2: Large Memory Models

DreamDPO: Aligning Text-to-3D Generation with Human Preferences via Direct Preference Optimization

Exploring the Limit of Outcome Reward for Learning Mathematical Reasoning

Training Language Models for Social Deduction with Multi-Agent Reinforcement Learning

The Hidden Life of Tokens: Reducing Hallucination of Large Vision-Language Models via Visual Information Steering

Efficient-vDiT: Efficient Video Diffusion Transformers With Attention Tile

ReasonFlux: Hierarchical LLM Reasoning via Scaling Thought Templates

SynthDetoxM: Modern LLMs are Few-Shot Parallel Detoxification Data Annotators

Matryoshka Quantization

Show-o Turbo: Towards Accelerated Unified Multimodal Understanding and Generation

CODESIM: Multi-Agent Code Generation and Problem Solving through Simulation-Driven Planning and Debugging

EVEv2: Improved Baselines for Encoder-Free Vision-Language Models

Embodied Red Teaming for Auditing Robotic Foundation Models

Forbidden Science: Dual-Use AI Challenge Benchmark and Scientific Refusal Tests

Lumina-Video: Efficient and Flexible Video Generation with Multi-scale Next-DiT