Yuanzhi Li

~Yuanzhi_Li1

论文总数

12.5

年均投稿

5.9

平均评分

接收情况13/25

会议分布

ICLR

COLM

ICML

NeurIPS

发表论文 (25 篇)

202511 篇

6.3

Interpretability of Language Models for Learning Hierarchical Structures

ICLR 2025Rejected

5.4

Understand Clean Generalization and Robust Overfitting in Adversarial Training from Two Theoretical Views: Representation Complexity and Training Dynamics

ICLR 2025Rejected

6.5

Physics of Language Models: Part 3.2, Knowledge Manipulation

ICLR 2025Poster

6.0

Adversarial Training Can Provably Improve Robustness: Theoretical Analysis of Feature Learning Process Under Structured Data

ICLR 2025Poster

6.1

On the Clean Generalization and Robust Overfitting in Adversarial Training from Two Theoretical Views: Representation Complexity and Training Dynamics

ICML 2025Poster

7.3

Physics of Language Models: Part 3.3, Knowledge Capacity Scaling Laws

ICLR 2025Spotlight

6.0

Physics of Language Models: Part 2.1, Grade-School Math and the Hidden Reasoning Process

ICLR 2025Poster

7.3

Understanding the Evolution of the Neural Tangent Kernel at the Edge of Stability

NeurIPS 2025Poster

6.8

Physics of Language Models: Part 2.2, How to Learn From Mistakes on Grade-School Math Problems

ICLR 2025Poster

4.3

Why Clean Generalization and Robust Overfitting Both Happen in Adversarial Training

ICLR 2024Rejected

6.5

Understanding Transferable Representation Learning and Zero-shot Transfer in CLIP

ICLR 2024Poster

3.8

How does overparametrization affect features?

ICLR 2024Rejected

6.8

SmartPlay : A Benchmark for LLMs as Intelligent Agents

ICLR 2024Poster

7.5

Role of Locality and Weight Sharing in Image-Based Tasks: A Sample Complexity Separation between CNNs, LCNs, and FCNs

ICLR 2024Spotlight

4.0

Positional Description Matters for Transformers Arithmetic

ICLR 2024Rejected

7.5

VisualWebBench: How Far Have Multimodal LLMs Evolved in Web Page Understanding and Grounding?

COLM 2024Poster

6.0

AgentKit: Structured LLM Reasoning with Dynamic Graphs

COLM 2024Poster

6.0

Textbooks Are All You Need

合作者 (20)

Yuanzhi Li

发表论文 (25 篇)

202511 篇

Interpretability of Language Models for Learning Hierarchical Structures

Understand Clean Generalization and Robust Overfitting in Adversarial Training from Two Theoretical Views: Representation Complexity and Training Dynamics

Physics of Language Models: Part 3.2, Knowledge Manipulation

Adversarial Training Can Provably Improve Robustness: Theoretical Analysis of Feature Learning Process Under Structured Data

On the Clean Generalization and Robust Overfitting in Adversarial Training from Two Theoretical Views: Representation Complexity and Training Dynamics

Physics of Language Models: Part 3.3, Knowledge Capacity Scaling Laws

Physics of Language Models: Part 2.1, Grade-School Math and the Hidden Reasoning Process

Understanding the Evolution of the Neural Tangent Kernel at the Edge of Stability

Physics of Language Models: Part 2.2, How to Learn From Mistakes on Grade-School Math Problems

Beyond Parameter Count: Implicit Bias in Soft Mixture of Experts

Mixture of Parrots: Experts improve memorization more than reasoning

202414 篇

TinyStories: How Small Can Language Models Be and Still Speak Coherent English

Simple mechanisms for representing, indexing and manipulating concepts

Knowledge Manipulation in Language Models (Part B)

Knowledge Storage and Extraction in Language Models (Part A)

How Language Models Learn Context-Free Grammars

Why Clean Generalization and Robust Overfitting Both Happen in Adversarial Training

Understanding Transferable Representation Learning and Zero-shot Transfer in CLIP

How does overparametrization affect features?

SmartPlay : A Benchmark for LLMs as Intelligent Agents

Role of Locality and Weight Sharing in Image-Based Tasks: A Sample Complexity Separation between CNNs, LCNs, and FCNs

Positional Description Matters for Transformers Arithmetic

VisualWebBench: How Far Have Multimodal LLMs Evolved in Web Page Understanding and Grounding?

AgentKit: Structured LLM Reasoning with Dynamic Graphs

Textbooks Are All You Need