Picture for Weijie J. Su

Weijie J. Su

A Peek into Token Bias: Large Language Models Are Not Yet Genuine Reasoners

Add code
Jun 16, 2024
Viaarxiv icon

Bridging the Gap: Rademacher Complexity in Robust and Standard Generalization

Jun 08, 2024
Viaarxiv icon

Tackling GenAI Copyright Issues: Originality Estimation and Genericization

Jun 05, 2024
Viaarxiv icon

Multi-Modal and Multi-Agent Systems Meet Rationality: A Survey

Add code
Jun 01, 2024
Viaarxiv icon

On the Algorithmic Bias of Aligning Large Language Models with RLHF: Preference Collapse and Matching Regularization

Add code
May 26, 2024
Viaarxiv icon

Neural Collapse Meets Differential Privacy: Curious Behaviors of NoisyGD with Near-perfect Representation Learning

May 16, 2024
Viaarxiv icon

An Economic Solution to Copyright Challenges of Generative AI

Add code
Apr 24, 2024
Figure 1 for An Economic Solution to Copyright Challenges of Generative AI
Figure 2 for An Economic Solution to Copyright Challenges of Generative AI
Figure 3 for An Economic Solution to Copyright Challenges of Generative AI
Figure 4 for An Economic Solution to Copyright Challenges of Generative AI
Viaarxiv icon

A Statistical Framework of Watermarks for Large Language Models: Pivot, Detection Efficiency and Optimal Rules

Apr 01, 2024
Viaarxiv icon

Provable Multi-Party Reinforcement Learning with Diverse Human Feedback

Mar 08, 2024
Viaarxiv icon

Unified Enhancement of Privacy Bounds for Mixture Mechanisms via $f$-Differential Privacy

Nov 01, 2023
Viaarxiv icon