PaperHub
影响力指数
96.88/100
前 0.2%
全站排名 #100
发表论文42 篇
平均评分5.8
年均产出14.0 篇/年

Percy Liang

Associate Professor@Stanford University·美国·OpenReview
研究方向

foundation models

7.5
13

Relative Scaling Laws for LLMs

ICLR 2026Desk Rejected
三作
7.5
15

Pre-training under infinite compute

ICLR 2026Oral
三作
6.5
14

Reinforcement Learning for Machine Learning Engineering Agents

ICLR 2026Poster
三作
6.5
10

WorldGym: World Model as An Environment for Policy Evaluation

ICLR 2026Poster
6.5
16

Cosmos Policy: Fine-Tuning Video Models for Visuomotor Control and Planning

ICLR 2026Poster
6.5
11

Comparing AI Agents to Cybersecurity Professionals in Real-World Penetration Testing

ICLR 2026Poster
5.5
21

Fantastic Pretraining Optimizers and Where to Find Them

ICLR 2026Poster
通讯
5.2
24

Holistic Agent Leaderboard: The Missing Infrastructure for AI Agent Evaluation

ICLR 2026Poster
5.0
22

MLE-Smith: Scaling MLE Tasks with Automated Multi-agent Pipeline

ICLR 2026Poster
4.0
16

SpecEval: Evaluating Model Adherence to Behavior Specifications

ICLR 2026Rejected
通讯
4.0
6

Curating High Quality Pretraining Data for Language Models via Compression Ratios

ICLR 2026Rejected
通讯
3.5
6

Scheduling data improves fine-tuning data efficiency

ICLR 2026Withdrawn
二作
3.5
16

UQ: Assessing Language Models on Unsolved Questions

ICLR 2026Rejected
3.3
4

AHELM: A Holistic Evaluation of Audio-Language Models

ICLR 2026Withdrawn
通讯
3.0
5

RoboReward: A Dataset and Benchmark for Vision-Language Reward Models in Robotics

ICLR 2026Withdrawn
3.0
5

Improving Dynamic Object Interactions in Text-to-Video Generation with AI Feedback

ICLR 2026Withdrawn
8.7
18

Cybench: A Framework for Evaluating Cybersecurity Capabilities and Risks of Language Models

ICLR 2025Oral
通讯
7.8
9

Eliciting Language Model Behaviors with Investigator Agents

ICML 2025Poster
7.5
15

AIR-BENCH 2024: A Safety Benchmark based on Regulation and Policies Specified Risk Categories

ICLR 2025Spotlight
7.3
19

Blackbox Model Provenance via Palimpsestic Membership Inference

NeurIPS 2025Spotlight
通讯
7.2
11

Reliable and Efficient Amortized Model-based Evaluation

ICML 2025Poster
三作
7.2
13

Language Models May Verbatim Complete Text They Were Not Explicitly Trained On

ICML 2025Spotlight
6.8
31

Audits Under Resource, Data, and Access Constraints: Scaling Laws For Less Discriminatory Alternatives

NeurIPS 2025Poster
6.5
26

Model Equality Testing: Which Model is this API Serving?

ICLR 2025Poster
二作
6.5
34

Reliable and Efficient Amortized Model-based Evaluation

ICLR 2025Rejected
三作
6.4
17

On the Entropy Calibration of Language Models

NeurIPS 2025Poster
三作
6.4
13

Independence Tests for Language Models

ICML 2025Spotlight
通讯
6.4
54

BioDiscoveryAgent: An AI Agent for Designing Genetic Perturbation Experiments

ICLR 2025Poster
6.3
29

AutoBencher: Towards Declarative Benchmark Construction

ICLR 2025Poster
6.1
12

Auditing Prompt Caching in Language Model APIs

ICML 2025Poster
6.0
20

Instruction Following without Instruction Tuning

ICLR 2025Rejected
通讯
6.0
15

Understanding Warmup-Stable-Decay Learning Rates: A River Valley Loss Landscape View

ICLR 2025Poster
5.3
23

Independence Tests for Language Models

ICLR 2025Rejected
通讯
4.8
24

VideoAgent: Self-Improving Video Generation

ICLR 2025Rejected
3.5
5

On the Entropy Calibration of Language Models

ICLR 2025Withdrawn
三作
合作者 (20)