AGI Lab

non-profit

AI & ML interests

None defined yet.

Papers

How LoRA Remembers? A Parametric Memory Law for LLM Finetuning

Rethinking Memory as Continuously Evolving Connectivity

View all Papers

AGI-LAB-HF 's papers 28

Submitted by

Ningyu Zhang

How LoRA Remembers? A Parametric Memory Law for LLM Finetuning

AGI-LAB-HF

Submitted by

Ningyu Zhang

Rethinking Memory as Continuously Evolving Connectivity

AGI-LAB-HF

2

Submitted by

Ningyu Zhang

MemTrace: Tracing and Attributing Errors in Large Language Model Memory Systems

AGI-LAB-HF

Submitted by

bowen xu

Long Live The Balance: Information Bottleneck Driven Tree-based Policy Optimization

AGI-LAB-HF

Submitted by

xiaochonglinghu

Enhancing Train-Free Infinite-Frame Generation for Consistent Long Videos

AGI-LAB-HF

Submitted by

shen

FashionChameleon: Towards Real-Time and Interactive Human-Garment Video Customization

AGI-LAB-HF

Submitted by

Jialiang Cheng

BEAM: Binary Expert Activation Masking for Dynamic Routing in MoE

AGI-LAB-HF

2

Submitted by

zhengrong yue

What Matters for Diffusion-Friendly Latent Manifold? Prior-Aligned Autoencoders for Latent Diffusion

AGI-LAB-HF

Submitted by

liutao

Continuous-Time Distribution Matching for Few-Step Diffusion Distillation

AGI-LAB-HF

Submitted by

xuguo

CoInteract: Physically-Consistent Human-Object Interaction Video Synthesis via Spatially-Structured Co-Generation

AGI-LAB-HF

Submitted by

xiaochonglinghu

Elucidating the SNR-t Bias of Diffusion Probabilistic Models

AGI-LAB-HF

Submitted by

xiaochonglinghu

Omni-WorldBench: Towards a Comprehensive Interaction-Centric Evaluation for World Models

AGI-LAB-HF