University of Southern California

university

https://www.usc.edu/

AI & ML interests

None defined yet.

Recent Activity

ZQ-Dev submitted a paper about 14 hours ago

Agent Mosaic: Recognizing Confidentiality Leaks in Agentic LLM Systems

donghyunli submitted a paper 12 days ago

SoloQ: Calibration-Free Quantization for Diffusion Language Models

yunblack submitted a paper 22 days ago

Prefill-Free Cross-Family KV Cache Transfer for Heterogeneous Multi-Agent LLMs

View all activity

Papers

Agent Mosaic: Recognizing Confidentiality Leaks in Agentic LLM Systems

3DCodeVerse: Exploring 3D Worlds as Code

View all Papers

UniversityofSouthernCalifornia 's papers 22

Submitted by

ZQ

Agent Mosaic: Recognizing Confidentiality Leaks in Agentic LLM Systems

UniversityofSouthernCalifornia

University of Southern California

Submitted by

Yipeng Gao

3DCodeVerse: Exploring 3D Worlds as Code

UniversityofSouthernCalifornia

University of Southern California

Submitted by

Donghyun Lee

SoloQ: Calibration-Free Quantization for Diffusion Language Models

UniversityofSouthernCalifornia

University of Southern California

Submitted by

Vincent-Daniel Yun

Prefill-Free Cross-Family KV Cache Transfer for Heterogeneous Multi-Agent LLMs

UniversityofSouthernCalifornia

University of Southern California

Submitted by

Fernandez

Rate-distortion optimization for full-reference image quality metrics via stochastic Hessian estimates

UniversityofSouthernCalifornia

University of Southern California

Submitted by

Yucheng Du

Recognition-Refusal Misalignment in LLMs: Why Models Answer Structurally Unanswerable Questions

UniversityofSouthernCalifornia

University of Southern California

Submitted by

Deqing Fu

Token-Level Off-Policy Learning for Faithful Generation Under Distribution Shift

UniversityofSouthernCalifornia

University of Southern California

2

Submitted by

Muzi Tao

An Exam for Active Observers

UniversityofSouthernCalifornia

University of Southern California

Submitted by

Donghyun Lee

KronQ: LLM Quantization via Kronecker-Factored Hessian

UniversityofSouthernCalifornia

University of Southern California

Submitted by

Yifan Jiang

AFFORDANCE20Q: Evaluating Affordance Reasoning from Physical Properties

UniversityofSouthernCalifornia

University of Southern California

Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs

UniversityofSouthernCalifornia

University of Southern California

Submitted by

Rishabh Agrawal

Reinforcement Learning from Rich Feedback with Distributional DAgger

UniversityofSouthernCalifornia

University of Southern California