PaperHub
影响力指数
99.27/100
前 0.1%
全站排名 #13
发表论文65 篇
平均评分5.7
年均产出21.7 篇/年

J Zico Kolter

Full Professor@Carnegie Mellon University·美国·OpenReview
研究方向

ai safety; deep learning; adversarial robustness; optimization; reinforcement learning; energy applications; convex optimization

7.0
19

Mamba-3: Improved Sequence Modeling using State Space Principles

ICLR 2026Oral
6.5
11

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

ICLR 2026Poster
5.3
18

Reevaluating Policy Gradient Methods for Imperfect-Information Games

ICLR 2026Poster
4.7
15

Accelerating Diffusion Planners in Offline RL via Reward-Aware Consistency Trajectory Distillation

ICLR 2026Poster
4.5
14

Looking beyond the next token

ICLR 2026Rejected
三作
4.5
16

Joint Distillation for Fast Likelihood Evaluation and Sampling in Flow-based Models

ICLR 2026Poster
4.5
18

Existing Adversarial LLM Unlearning Evaluations Are Inconclusive

ICLR 2026Rejected
通讯
4.0
11

Not All Rollouts are Useful: Down-Sampling Rollouts in LLM Reinforcement Learning

ICLR 2026Withdrawn
通讯
3.0
5

Towards Understanding the Transferability of Adversarial Suffixes in Large Language Models

ICLR 2026Withdrawn
8.0
15

Context-Parametric Inversion: Why Instruction Finetuning May Not Actually Improve Context Reliance

ICLR 2025Oral
三作
7.8
27

Mean Flows for One-step Generative Modeling

NeurIPS 2025Oral
7.8
15

Training a Generally Curious Agent

ICML 2025Oral
7.3
16

Predicting the Performance of Black-box Language Models with Follow-up Queries

NeurIPS 2025Poster
三作
7.0
23

Understanding Optimization in Deep Learning with Central Flows

ICLR 2025Poster
7.0
18

Weight ensembling improves reasoning in language models

COLM 2025Poster
6.8
28

Safety Pretraining: Toward the Next Generation of Safe AI

NeurIPS 2025Poster
通讯
6.8
8

Consistency Models Made Easy

ICLR 2025Poster
通讯
6.8
19

AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents

ICLR 2025Poster
6.4
16

Antidistillation Sampling

NeurIPS 2025Poster
通讯
6.0
34

Eliciting Black-Box Representations from LLMs through Self-Queries

ICLR 2025Rejected
三作
6.0
19

Compute-Optimal LLMs Provably Generalize Better with Scale

ICLR 2025Poster
5.8
41

Inference Optimal VLMs Need Fewer Visual Tokens and More Parameters

ICLR 2025Poster
通讯
5.5
13

Idiosyncrasies in Large Language Models

ICML 2025Poster
5.5
24

Adaptive Data Optimization: Dynamic Sample Selection with Scaling Laws

ICLR 2025Poster
通讯
5.5
13

Contextures: Representations from Contexts

ICML 2025Poster
5.4
25

Bayesian Neural Networks with Domain Knowledge Priors

ICLR 2025Rejected
通讯
5.3
13

The Promises and Pitfalls of Language Models for Structured Numerical Data

ICLR 2025Rejected
5.3
20

Finetuning CLIP to Reason about Pairwise Differences

ICLR 2025Rejected
通讯
5.0
14

Data-scarce distillation for large-scale vision language models

ICLR 2025Rejected
通讯
5.0
30

Automated Black-box Prompt Engineering for Personalized Text-to-Image Generation

ICLR 2025Rejected
通讯
4.9
15

Unnatural Languages Are Not Bugs but Features for LLMs

ICML 2025Poster
4.6
18

DEQ-MPC : Deep Equilibrium Model Predictive Control

ICLR 2025Rejected
通讯
4.5
22

Mimetic Initialization Helps State Space Models Learn to Recall

ICLR 2025Rejected
三作
4.0
9

HyperCLIP: Adapting Vision-Language models with Hypernetworks

ICLR 2025Rejected
通讯
3.0
13

Prompt Recovery for Image Generation Models: A Comparative Study of Discrete Optimizers

ICLR 2025Rejected
通讯
合作者 (20)