Publications

2025

• Establishing best practices for building rigorous agentic benchmarks. Yuxuan Zhu, Tengjun Jin, Yada Pruksachatkun, Andy Zhang, Shu Liu, Sasha Cui, Sayash Kapoor, Shayne Longpre, Kevin Meng, Rebecca Weiss, Fazl Barez, Rahul Gupta, Jwala Dhamala, Jacob Merizian, Mario Giulianelli, Harry Coppock, Cozmin Ududec, Jasjeet Sekhon, Jacob Steinhardt, Antony Kellerman, Sarah Schwettmann, Matei Zaharia, Ion Stoica, Percy Liang, Daniel Kang. arXiv, 2025.

  @article{zhu2025establishing,
    author = {Yuxuan Zhu and Tengjun Jin and Yada Pruksachatkun and Andy Zhang and Shu Liu and Sasha Cui and Sayash Kapoor and Shayne Longpre and Kevin Meng and Rebecca Weiss and Fazl Barez and Rahul Gupta and Jwala Dhamala and Jacob Merizian and Mario Giulianelli and Harry Coppock and Cozmin Ududec and Jasjeet Sekhon and Jacob Steinhardt and Antony Kellerman and Sarah Schwettmann and Matei Zaharia and Ion Stoica and Percy Liang and Daniel Kang},
    journal = {arXiv},
    title = {Establishing Best Practices for Building Rigorous Agentic Benchmarks},
    year = {2025},
  }
  

  [bib] [paper]
  
• Understanding in-context learning of addition via activation subspaces. Xinyan Hu, Kayo Yin, Michael I. Jordan, Jacob Steinhardt, Lijie Chen. arXiv, 2025.

  @article{hu2025understanding,
    author = {Xinyan Hu and Kayo Yin and Michael I. Jordan and Jacob Steinhardt and Lijie Chen},
    journal = {arXiv},
    title = {Understanding In-context Learning of Addition via Activation Subspaces},
    year = {2025},
  }
  

  [bib] [paper]
  
• Uncovering gaps in how humans and LLMs interpret subjective language. Erik Jones, Arjun Patrawala, Jacob Steinhardt. International Conference on Learning Representations (ICLR), 2025. Spotlight presentation.

  @inproceedings{jones2025uncovering,
    author = {Erik Jones and Arjun Patrawala and Jacob Steinhardt},
    booktitle = {International Conference on Learning Representations (ICLR)},
    title = {Uncovering Gaps in How Humans and {LLM}s Interpret Subjective Language},
    year = {2025},
  }
  

  [bib] [paper]
  
• Eliciting language model behaviors with investigator agents. Xiang Lisa Li, Neil Chowdhury, Daniel D. Johnson, Tatsunori Hashimoto, Percy Liang, Sarah Schwettmann, Jacob Steinhardt. arXiv, 2025.

  @article{xiang2025eliciting,
    author = {Xiang Lisa Li and Neil Chowdhury and Daniel D. Johnson and Tatsunori Hashimoto and Percy Liang and Sarah Schwettmann and Jacob Steinhardt},
    journal = {arXiv},
    title = {Eliciting Language Model Behaviors with Investigator Agents},
    year = {2025},
  }
  

  [bib] [paper]
  
• Iterative label refinement matters more than preference optimization under weak supervision. Yaowen Ye, Cassidy Laidlaw, Jacob Steinhardt. arXiv, 2025.

  @article{ye2025iterative,
    author = {Yaowen Ye and Cassidy Laidlaw and Jacob Steinhardt},
    journal = {arXiv},
    title = {Iterative Label Refinement Matters More than Preference Optimization under Weak Supervision},
    year = {2025},
  }
  

  [bib] [paper]
  

2024

• LatentQA: teaching LLMs to decode activations into natural language. Alexander Pan, Lijie Chen, Jacob Steinhardt. arXiv, 2024.

  @article{pan2024latent,
    author = {Alexander Pan and Lijie Chen and Jacob Steinhardt},
    journal = {arXiv},
    title = {LatentQA: Teaching {LLMs} to Decode Activations Into Natural Language},
    year = {2024},
  }
  

  [bib] [paper]
  
• Extractive structures learned in pretraining enable generalization on finetuned facts. Jiahai Feng, Stuart Russell, Jacob Steinhardt. arXiv, 2024.

  @article{feng2024extractive,
    author = {Jiahai Feng and Stuart Russell and Jacob Steinhardt},
    journal = {arXiv},
    title = {Extractive Structures Learned in Pretraining Enable Generalization on Finetuned Facts},
    year = {2024},
  }
  

  [bib] [paper]
  
• What do learning dynamics reveal about generalization in LLM reasoning? Katie Kang, Amrith Setlur, Dibya Ghosh, Jacob Steinhardt, Claire Tomlin, Sergey Levine, Aviral Kumar. arXiv, 2024.

  @article{kang2024dynamics,
    author = {Katie Kang and Amrith Setlur and Dibya Ghosh and Jacob Steinhardt and Claire Tomlin and Sergey Levine and Aviral Kumar},
    journal = {arXiv},
    title = {What Do Learning Dynamics Reveal About Generalization in {LLM} Reasoning?},
    year = {2024},
  }
  

  [bib] [paper]
  
• VibeCheck: discover and quantify qualitative differences in large language models. Lisa Dunlap, Krishna Mandal, Trevor Darrell, Jacob Steinhardt, Joseph E Gonzalez. arXiv, 2024.

  @article{dunlap2024vibecheck,
    author = {Lisa Dunlap and Krishna Mandal and Trevor Darrell and Jacob Steinhardt and Joseph E Gonzalez},
    journal = {arXiv},
    title = {VibeCheck: Discover and Quantify Qualitative Differences in Large Language Models},
    year = {2024},
  }
  

  [bib] [paper]
  
• Explaining datasets in words: statistical models with natural language parameters. Ruiqi Zhong, Heng Wang, Dan Klein, Jacob Steinhardt. arXiv, 2024.

  @article{zhong2024explaining,
    author = {Ruiqi Zhong and Heng Wang and Dan Klein and Jacob Steinhardt},
    journal = {arXiv},
    title = {Explaining Datasets in Words: Statistical Models with Natural Language Parameters},
    year = {2024},
  }
  

  [bib] [paper]
  
• Safety vs. performance: how multi-objective learning reduces barriers to market entry. Meena Jagadeesan, Michael I. Jordan, Jacob Steinhardt. arXiv, 2024.

  @article{jagadeesan2024safety,
    author = {Meena Jagadeesan and Michael I. Jordan and Jacob Steinhardt},
    journal = {arXiv},
    title = {Safety vs. Performance: How Multi-Objective Learning Reduces Barriers to Market Entry},
    year = {2024},
  }
  

  [bib] [paper]
  
• Feedback loops with language models drive in-context reward hacking. Alexander Pan, Erik Jones, Meena Jagadeesan, Jacob Steinhardt. International Conference on Machine Learning (ICML), 2024.

  @inproceedings{pan2024feedback,
    author = {Alexander Pan and Erik Jones and Meena Jagadeesan and Jacob Steinhardt},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Feedback Loops With Language Models Drive In-Context Reward Hacking},
    year = {2024},
  }
  

  [bib] [paper]
  
• Do models explain themselves? counterfactual simulatability of natural language explanations. Yanda Chen, Ruiqi Zhong, Narutatsu Ri, Chen Zhao, He He, Jacob Steinhardt, Zhou Yu, Kathleen McKeown. International Conference on Machine Learning (ICML), 2024. Spotlight presentation.

  @inproceedings{chen2024models,
    author = {Yanda Chen and Ruiqi Zhong and Narutatsu Ri and Chen Zhao and He He and Jacob Steinhardt and Zhou Yu and Kathleen McKeown},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations},
    year = {2024},
  }
  

  [bib] [paper]
  
• Monitoring latent world states in language models with propositional probes. Jiahai Feng, Stuart Russell, Jacob Steinhardt. arXiv, 2024.

  @article{feng2024monitoring,
    author = {Jiahai Feng and Stuart Russell and Jacob Steinhardt},
    journal = {arXiv},
    title = {Monitoring Latent World States in Language Models with Propositional Probes},
    year = {2024},
  }
  

  [bib] [paper]
  
• Adversaries can misuse combinations of safe models. Erik Jones, Anca Dragan, Jacob Steinhardt. arXiv, 2024.

  @article{jones2024adversaries,
    author = {Erik Jones and Anca Dragan and Jacob Steinhardt},
    journal = {arXiv},
    title = {Adversaries Can Misuse Combinations of Safe Models},
    year = {2024},
  }
  

  [bib] [paper]
  
• Interpreting the second-order effects of neurons in CLIP. Yossi Gandelsman, Alexei A. Efros, Jacob Steinhardt. arXiv, 2024.

  @article{gandelsman2024interpreting,
    author = {Yossi Gandelsman and Alexei A. Efros and Jacob Steinhardt},
    journal = {arXiv},
    title = {Interpreting the Second-Order Effects of Neurons in {CLIP}},
    year = {2024},
  }
  

  [bib] [paper]
  
• Overthinking the truth: understanding how language models process false demonstrations. Danny Halawi, Jean-Stanislas Denain, Jacob Steinhardt. International Conference on Learning Representations (ICLR), 2024. Spotlight presentation.

  @inproceedings{halawi2024overthinking,
    author = {Danny Halawi and Jean-Stanislas Denain and Jacob Steinhardt},
    booktitle = {International Conference on Learning Representations (ICLR)},
    title = {Overthinking the Truth: Understanding how Language Models Process False Demonstrations},
    year = {2024},
  }
  

  [bib] [paper]
  
• Protein language models are biased by unequal sequence sampling across the tree of life. Frances Ding, Jacob Steinhardt. arXiv, 2024.

  @article{ding2024protein,
    author = {Frances Ding and Jacob Steinhardt},
    journal = {arXiv},
    title = {Protein language models are biased by unequal sequence sampling across the tree of life},
    year = {2024},
  }
  

  [bib] [paper]
  
• Approaching human-level forecasting with language models. Danny Halawi, Fred Zhang, Chen Yueh-Han, Jacob Steinhardt. arXiv, 2024.

  @article{halawi2024approaching,
    author = {Danny Halawi and Fred Zhang and Chen Yueh-Han and Jacob Steinhardt},
    journal = {arXiv},
    title = {Approaching Human-Level Forecasting with Language Models},
    year = {2024},
  }
  

  [bib] [paper]
  
• Describing differences in image sets with natural language. Lisa Dunlap, Yuhui Zhang, Xiaohan Wang, Ruiqi Zhong, Trevor Darrell, Jacob Steinhardt, Joseph E. Gonzalez, Serena Yeung-Levy. Computer Vision and Pattern Recognition (CVPR), 2024. Oral presentation.

  @inproceedings{dunlap2024describing,
    author = {Lisa Dunlap and Yuhui Zhang and Xiaohan Wang and Ruiqi Zhong and Trevor Darrell and Jacob Steinhardt and Joseph E. Gonzalez and Serena Yeung-Levy},
    booktitle = {Computer Vision and Pattern Recognition (CVPR)},
    title = {Describing Differences in Image Sets with Natural Language},
    year = {2024},
  }
  

  [bib] [paper]
  
• How do language models bind entities in context? Jiahai Feng, Jacob Steinhardt. International Conference on Learning Representations (ICLR), 2024.

  @inproceedings{feng2024entities,
    author = {Jiahai Feng and Jacob Steinhardt},
    booktitle = {International Conference on Learning Representations (ICLR)},
    title = {How do Language Models Bind Entities in Context?},
    year = {2024},
  }
  

  [bib] [paper]
  
• Interpreting CLIP's image representation via text-based decomposition. Yossi Gandelsman, Alexei A. Efros, Jacob Steinhardt. International Conference on Learning Representations (ICLR), 2024. Oral presentation.

  @inproceedings{gandelsman2024mass,
    author = {Yossi Gandelsman and Alexei A. Efros and Jacob Steinhardt},
    booktitle = {International Conference on Learning Representations (ICLR)},
    title = {Interpreting {CLIP}'s Image Representation via Text-Based Decomposition},
    year = {2024},
  }
  

  [bib] [paper]
  

2023

• Learning equilibria in matching markets from bandit feedback. Meena Jagadeesan, Alexander Wei, Yixin Wang, Michael I. Jordan, Jacob Steinhardt. Journal of the ACM (JACM), 2023.

  @article{jagadeesan2023learning,
    author = {Meena Jagadeesan and Alexander Wei and Yixin Wang and Michael I. Jordan and Jacob Steinhardt},
    journal = {Journal of the ACM (JACM)},
    title = {Learning Equilibria in Matching Markets from Bandit Feedback},
    volume = {70},
    year = {2023},
  }
  

  [bib] [paper]
  
• Mass-Producing failures of multimodal systems with language models. Shengbang Tong, Erik Jones, Jacob Steinhardt. Advances in Neural Information Processing Systems (NeurIPS), 2023.

  @inproceedings{tong2023mass,
    author = {Shengbang Tong and Erik Jones and Jacob Steinhardt},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Mass-Producing Failures of Multimodal Systems with Language Models},
    year = {2023},
  }
  

  [bib] [paper]
  
• Jailbroken: how does LLM safety training fail? Alexander Wei, Nika Haghtalab, Jacob Steinhardt. Advances in Neural Information Processing Systems (NeurIPS), 2023. Oral presentation.

  @inproceedings{wei2023jailbroken,
    author = {Alexander Wei and Nika Haghtalab and Jacob Steinhardt},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Jailbroken: How Does {LLM} Safety Training Fail?},
    year = {2023},
  }
  

  [bib] [paper]
  
• Supply-Side equilibria in recommender systems. Meena Jagadeesan, Nikhil Garg, Jacob Steinhardt. Advances in Neural Information Processing Systems (NeurIPS), 2023.

  @inproceedings{jagadeesan2023supply,
    author = {Meena Jagadeesan and Nikhil Garg and Jacob Steinhardt},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Supply-Side Equilibria in Recommender Systems},
    year = {2023},
  }
  

  [bib] [paper]
  
• Improved bayes risk can yield reduced social welfare under competition. Meena Jagadeesan, Michael I. Jordan, Jacob Steinhardt, Nika Haghtalab. Advances in Neural Information Processing Systems (NeurIPS), 2023.

  @inproceedings{jagadeesan2023improved,
    author = {Meena Jagadeesan and Michael I. Jordan and Jacob Steinhardt and Nika Haghtalab},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Improved Bayes Risk Can Yield Reduced Social Welfare Under Competition},
    year = {2023},
  }
  

  [bib] [paper]
  
• Are neurons actually collapsed? on the fine-grained structure in neural representations. Yongyi Yang, Jacob Steinhardt, Wei Hu. International Conference on Machine Learning (ICML), 2023.

  @inproceedings{yang2023neurons,
    author = {Yongyi Yang and Jacob Steinhardt and Wei Hu},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Are Neurons Actually Collapsed? On the Fine-Grained Structure in Neural Representations},
    year = {2023},
  }
  

  [bib] [paper]
  
• Automatically auditing large language models via discrete optimization. Erik Jones, Anca Dragan, Aditi Raghunathan, Jacob Steinhardt. International Conference on Machine Learning (ICML), 2023.

  @inproceedings{jones2023automatically,
    author = {Erik Jones and Anca Dragan and Aditi Raghunathan and Jacob Steinhardt},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Automatically Auditing Large Language Models via Discrete Optimization},
    year = {2023},
  }
  

  [bib] [paper]
  
• Interpretability in the wild: a circuit for indirect object identification in GPT-2 small. Kevin Wang, Alexandre Variengien, Arthur Conmy, Buck Shlegeris, Jacob Steinhardt. International Conference on Learning Representations (ICLR), 2023.

  @inproceedings{wang2023interpretability,
    author = {Kevin Wang and Alexandre Variengien and Arthur Conmy and Buck Shlegeris and Jacob Steinhardt},
    booktitle = {International Conference on Learning Representations (ICLR)},
    title = {Interpretability in the Wild: a Circuit for Indirect Object Identification in {GPT}-2 small},
    year = {2023},
  }
  

  [bib] [paper]
  
• Progress measures for grokking via mechanistic interpretability. Neel Nanda, Lawrence Chan, Tom Liberum, Jess Smith, Jacob Steinhardt. International Conference on Learning Representations (ICLR), 2023. Oral presentation.

  @inproceedings{nanda2023progress,
    author = {Neel Nanda and Lawrence Chan and Tom Liberum and Jess Smith and Jacob Steinhardt},
    booktitle = {International Conference on Learning Representations (ICLR)},
    title = {Progress measures for grokking via mechanistic interpretability},
    year = {2023},
  }
  

  [bib] [paper]
  
• Incentivizing high-quality content in online recommender systems. Xinyan Hu, Meena Jagadeesan, Michael I. Jordan, Jacob Steinhardt. arXiv, 2023.

  @article{hu2023incentivizing,
    author = {Xinyan Hu and Meena Jagadeesan and Michael I. Jordan and Jacob Steinhardt},
    journal = {arXiv},
    title = {Incentivizing High-Quality Content in Online Recommender Systems},
    year = {2023},
  }
  

  [bib] [paper]
  
• Eliciting latent predictions from transformers with the tuned lens. Nora Belrose, Zach Furman, Logan Smith, Danny Halawi, Igor Ostrovsky, Lev McKinney, Stella Biderman, Jacob Steinhardt. arXiv, 2023.

  @article{belrose2023eliciting,
    author = {Nora Belrose and Zach Furman and Logan Smith and Danny Halawi and Igor Ostrovsky and Lev McKinney and Stella Biderman and Jacob Steinhardt},
    journal = {arXiv},
    title = {Eliciting Latent Predictions from Transformers with the Tuned Lens},
    year = {2023},
  }
  

  [bib] [paper]
  
• Goal driven discovery of distributional differences via language descriptions. Ruiqi Zhong, Peter Zhang, Steve Li, Jinwoo Ahn, Dan Klein, Jacob Steinhardt. Advances in Neural Information Processing Systems (NeurIPS), 2023.

  @inproceedings{zhong2023goal,
    author = {Ruiqi Zhong and Peter Zhang and Steve Li and Jinwoo Ahn and Dan Klein and Jacob Steinhardt},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Goal Driven Discovery of Distributional Differences via Language Descriptions},
    year = {2023},
  }
  

  [bib] [paper]
  
• Discovering latent knowledge in language models without supervision. Collin Burns, Haotian Ye, Dan Klein, Jacob Steinhardt. International Conference on Learning Representations (ICLR), 2023.

  @inproceedings{burns2023discovering,
    author = {Collin Burns and Haotian Ye and Dan Klein and Jacob Steinhardt},
    booktitle = {International Conference on Learning Representations (ICLR)},
    title = {Discovering latent knowledge in language models without supervision},
    year = {2023},
  }
  

  [bib] [paper]
  

2022

• Generalized resilience and robust statistics. Banghua Zhu, Jiantao Jiao, Jacob Steinhardt. Annals of Statistics, 2022.

  @article{zhu2022generalized,
    author = {Banghua Zhu and Jiantao Jiao and Jacob Steinhardt},
    journal = {Annals of Statistics},
    title = {Generalized Resilience and Robust Statistics},
    volume = {50},
    year = {2022},
  }
  

  [bib] [paper] [talk]
  
• How would the viewer feel? estimating wellbeing from video scenarios. Mantas Mazeika, Eric Tang, Andy Zou, Steven Basart, Jun Shern Chan, Dawn Song, David Forsyth, Jacob Steinhardt, Dan Hendrycks. Advances in Neural Information Processing Systems (NeurIPS), 2022.

  @inproceedings{mazeika2022viewer,
    author = {Mantas Mazeika and Eric Tang and Andy Zou and Steven Basart and Jun Shern Chan and Dawn Song and David Forsyth and Jacob Steinhardt and Dan Hendrycks},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {How Would The Viewer Feel? Estimating Wellbeing From Video Scenarios},
    year = {2022},
  }
  

  [bib] [paper]
  
• Capturing failures of large language models via human cognitive biases. Erik Jones, Jacob Steinhardt. Advances in Neural Information Processing Systems (NeurIPS), 2022.

  @inproceedings{jones2022capturing,
    author = {Erik Jones and Jacob Steinhardt},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Capturing Failures of Large Language Models via Human Cognitive Biases},
    year = {2022},
  }
  

  [bib] [paper]
  
• A3D: studying pretrained representations with programmable datasets. Ye Wang, Norman Mu, Daniele Grandi, Nicolas Savva, Jacob Steinhardt. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2022.

  @inproceedings{wang2022ad3,
    author = {Ye Wang and Norman Mu and Daniele Grandi and Nicolas Savva and Jacob Steinhardt},
    booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
    title = {A3D: Studying Pretrained Representations With Programmable Datasets},
    year = {2022},
  }
  

  [bib] [paper]
  
• Forecasting future world events with neural networks. Andy Zou, Tristan Xiao, Ryan Jia, Joe Kwon, Mantas Mazeika, Richard Li, Dawn Song, Jacob Steinhardt, Owain Evans, Dan Hendrycks. Advances in Neural Information Processing Systems (NeurIPS), 2022.

  @inproceedings{zou2022forecasting,
    author = {Andy Zou and Tristan Xiao and Ryan Jia and Joe Kwon and Mantas Mazeika and Richard Li and Dawn Song and Jacob Steinhardt and Owain Evans and Dan Hendrycks},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Forecasting Future World Events with Neural Networks},
    year = {2022},
  }
  

  [bib] [paper]
  
• Auditing visualizations: transparency methods struggle to detect anomalous behavior. Jean-Stanislas Denain, Jacob Steinhardt. arXiv, 2022.

  @article{denain2022auditing,
    author = {Jean-Stanislas Denain and Jacob Steinhardt},
    journal = {arXiv},
    title = {Auditing Visualizations: Transparency Methods Struggle to Detect Anomalous Behavior},
    year = {2022},
  }
  

  [bib] [paper]
  
• More than a toy: random matrix models predict how real-world neural representations generalize. Alexander Wei, Wei Hu, Jacob Steinhardt. International Conference on Machine Learning (ICML), 2022.

  @inproceedings{wei2022more,
    author = {Alexander Wei and Wei Hu and Jacob Steinhardt},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {More Than a Toy: Random Matrix Models Predict How Real-World Neural Representations Generalize},
    year = {2022},
  }
  

  [bib] [paper]
  
• Predicting out-of-distribution error with the projection norm. Yaodong Yu, Zitong Yang, Alexander Wei, Yi Ma, Jacob Steinhardt. International Conference on Machine Learning (ICML), 2022.

  @inproceedings{yu2022predicting,
    author = {Yaodong Yu and Zitong Yang and Alexander Wei and Yi Ma and Jacob Steinhardt},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Predicting Out-of-Distribution Error with the Projection Norm},
    year = {2022},
  }
  

  [bib] [paper]
  
• Describing differences between text distributions with natural language. Ruiqi Zhong, Charlie Snell, Dan Klein, Jacob Steinhardt. International Conference on Machine Learning (ICML), 2022.

  @inproceedings{zhong2022summarizing,
    author = {Ruiqi Zhong and Charlie Snell and Dan Klein and Jacob Steinhardt},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Describing differences between text distributions with natural language},
    year = {2022},
  }
  

  [bib] [paper]
  
• The effects of reward misspecification: mapping and mitigating misaligned models. Alexander Pan, Kush Bhatia, Jacob Steinhardt. International Conference on Learning Representations (ICLR), 2022.

  @inproceedings{pan2022effects,
    author = {Alexander Pan and Kush Bhatia and Jacob Steinhardt},
    booktitle = {International Conference on Learning Representations (ICLR)},
    title = {The Effects of Reward Misspecification: Mapping and Mitigating Misaligned Models},
    year = {2022},
  }
  

  [bib] [paper]
  
• PixMix: dreamlike pictures comprehensively improve safety measures. Dan Hendrycks, Andy Zou, Mantas Mazeika, Leonard Tang, Dawn Song, Jacob Steinhardt. Computer Vision and Pattern Recognition (CVPR), 2022.

  @inproceedings{hendrycks2022pixmix,
    author = {Dan Hendrycks and Andy Zou and Mantas Mazeika and Leonard Tang and Dawn Song and Jacob Steinhardt},
    booktitle = {Computer Vision and Pattern Recognition (CVPR)},
    title = {PixMix: Dreamlike Pictures Comprehensively Improve Safety Measures},
    year = {2022},
  }
  

  [bib] [paper] [code]
  
• Stronger data poisoning attacks break data sanitization defenses. Pang Wei Koh, Jacob Steinhardt, Percy Liang. Machine Learning, 2022.

  @article{koh2022stronger,
    author = {Pang Wei Koh and Jacob Steinhardt and Percy Liang},
    journal = {Machine Learning},
    number = {1},
    pages = {1--47},
    title = {Stronger data poisoning attacks break data sanitization defenses},
    volume = {111},
    year = {2022},
  }
  

  [bib] [paper] [code]
  
• Scaling out-of-distribution detection for real-world settings. Dan Hendrycks, Steven Basart, Mantas Mazeika, Mohammadreza Mostajabi, Jacob Steinhardt, Dawn Song. International Conference on Machine Learning (ICML), 2022.

  @inproceedings{hendrycks2022scaling,
    author = {Dan Hendrycks and Steven Basart and Mantas Mazeika and Mohammadreza Mostajabi and Jacob Steinhardt and Dawn Song},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Scaling out-of-distribution detection for real-world settings},
    year = {2022},
  }
  

  [bib] [paper] [data]
  

2021

• The effect of model size on worst-group generalization. Alan Le Pham, Eunice Chan, Vikranth Srivatsa, Dhruba Ghosh, Yaoqing Yang, Yaodong Yu, Ruiqi Zhong, Joseph E Gonzalez, Jacob Steinhardt. Advances in Neural Information Processing Systems Workshop (NeurIPS Workshop), 2021.

  @inproceedings{le2021effect,
    author = {Alan Le Pham and Eunice Chan and Vikranth Srivatsa and Dhruba Ghosh and Yaoqing Yang and Yaodong Yu and Ruiqi Zhong and Joseph E Gonzalez and Jacob Steinhardt},
    booktitle = {Advances in Neural Information Processing Systems Workshop (NeurIPS Workshop)},
    title = {The Effect of Model Size on Worst-Group Generalization},
    year = {2021},
  }
  

  [bib] [paper]
  
• What would jiminy cricket do? Towards agents that behave morally. Dan Hendrycks, Mantas Mazeika, Andy Zou, Sahil Patel, Christine Zhu, Jesus Navarro, Dawn Song, Bo Li, Jacob Steinhardt. Advances in Neural Information Processing Systems (Datasets and Benchmarks Track), 2021.

  @inproceedings{hendrycks2021jiminy,
    author = {Dan Hendrycks and Mantas Mazeika and Andy Zou and Sahil Patel and Christine Zhu and Jesus Navarro and Dawn Song and Bo Li and Jacob Steinhardt},
    booktitle = {Advances in Neural Information Processing Systems (Datasets and Benchmarks Track)},
    title = {What Would Jiminy Cricket Do? {T}owards Agents That Behave Morally},
    year = {2021},
  }
  

  [bib] [paper] [code]
  
• Unsolved problems in ML safety. Dan Hendrycks, Nicholas Carlini, John Schulman, Jacob Steinhardt. arXiv, 2021.

  @article{hendrycks2021unsolved,
    author = {Dan Hendrycks and Nicholas Carlini and John Schulman and Jacob Steinhardt},
    journal = {arXiv},
    title = {Unsolved Problems in {ML} Safety},
    year = {2021},
  }
  

  [bib] [paper] [blog post]
  
• Grounding representation similarity with statistical testing. Frances Ding, Jean-Stanislas Denain, Jacob Steinhardt. Advances in Neural Information Processing Systems (NeurIPS), 2021.

  @inproceedings{ding2021grounding,
    author = {Frances Ding and Jean-Stanislas Denain and Jacob Steinhardt},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Grounding Representation Similarity with Statistical Testing},
    year = {2021},
  }
  

  [bib] [paper]
  
• Constructing and adjusting estimates for household transmission of SARS-CoV-2 from prior studies, widespread-testing and contact-tracing data. Mihaela Curmei*, Andrew Ilyas*, Owain Evans, Jacob Steinhardt. International Journal of Epidemiology, 2021.

  @article{curmei2021estimating,
    author = {Mihaela Curmei and Andrew Ilyas and Owain Evans and Jacob Steinhardt},
    journal = {International Journal of Epidemiology},
    title = {Constructing and adjusting estimates for household transmission of {SARS}-{CoV}-2 from prior studies, widespread-testing and contact-tracing data},
    year = {2021},
  }
  

  [bib] [preprint] [journal]
  
• Robust estimation via generalized quasi-gradients. Banghua Zhu, Jiantao Jiao, Jacob Steinhardt. Information and Inference: A Journal of the IMA, 2021.

  @article{zhu2021robust,
    author = {Banghua Zhu and Jiantao Jiao and Jacob Steinhardt},
    journal = {Information and Inference: A Journal of the IMA},
    title = {Robust Estimation via Generalized Quasi-gradients},
    year = {2021},
  }
  

  [bib] [preprint] [journal]
  
• Measuring coding challenge competence with APPS. Dan Hendrycks, Steven Basart, Saurav Kadavath, Mantas Mazeika, Akul Arora, Ethan Guo, Collin Burns, Samir Puranik, Horace He, Dawn Song, Jacob Steinhardt. Advances in Neural Information Processing Systems (Datasets and Benchmarks Track), 2021.

  @inproceedings{hendrycksapps2021,
    author = {Dan Hendrycks and Steven Basart and Saurav Kadavath and Mantas Mazeika and Akul Arora and Ethan Guo and Collin Burns and Samir Puranik and Horace He and Dawn Song and Jacob Steinhardt},
    booktitle = {Advances in Neural Information Processing Systems (Datasets and Benchmarks Track)},
    title = {Measuring Coding Challenge Competence With {APPS}},
    year = {2021},
  }
  

  [bib] [paper] [data]
  
• Are larger pretrained language models uniformly better? Comparing performance at the instance level. Ruiqi Zhong, Dhruba Ghosh, Dan Klein, Jacob Steinhardt. Findings of the Association for Computational Linguistics (Findings of ACL), 2021.

  @inproceedings{zhong2021larger,
    author = {Ruiqi Zhong and Dhruba Ghosh and Dan Klein and Jacob Steinhardt},
    booktitle = {Findings of the Association for Computational Linguistics (Findings of ACL)},
    title = {Are Larger Pretrained Language Models Uniformly Better? {C}omparing Performance at the Instance Level},
    year = {2021},
  }
  

  [bib] [paper]
  
• Agnostic learning with unknown utilities. Kush Bhatia, Peter L. Bartlett, Anca D. Dragan, Jacob Steinhardt. Innovations in Theoretical Computer Science (ITCS), 2021.

  @inproceedings{bhatia2021agnostic,
    author = {Kush Bhatia and Peter L. Bartlett and Anca D. Dragan and Jacob Steinhardt},
    booktitle = {Innovations in Theoretical Computer Science (ITCS)},
    title = {Agnostic Learning with Unknown Utilities},
    year = {2021},
  }
  

  [bib] [paper] [talk]
  
• Measuring massive multitask language understanding. Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, Jacob Steinhardt. International Conference on Learning Representations (ICLR), 2021.

  @inproceedings{hendrycks2021measuring,
    author = {Dan Hendrycks and Collin Burns and Steven Basart and Andy Zou and Mantas Mazeika and Dawn Song and Jacob Steinhardt},
    booktitle = {International Conference on Learning Representations (ICLR)},
    title = {Measuring Massive Multitask Language Understanding},
    year = {2021},
  }
  

  [bib] [paper] [data]
  
• Aligning AI with shared human values. Dan Hendrycks, Collin Burns, Steven Basart, Andrew Critch, Jerry Li, Dawn Song, Jacob Steinhardt. International Conference on Learning Representations (ICLR), 2021.

  @inproceedings{hendrycks2021aligning,
    author = {Dan Hendrycks and Collin Burns and Steven Basart and Andrew Critch and Jerry Li and Dawn Song and Jacob Steinhardt},
    booktitle = {International Conference on Learning Representations (ICLR)},
    title = {Aligning {AI} With Shared Human Values},
    year = {2021},
  }
  

  [bib] [paper] [data]
  
• The many faces of robustness: a critical analysis of out-of-distribution generalization. Dan Hendrycks, Steven Basart, Norman Mu, Saurav Kadavath, Frank Wang, Evan Dorundo, Rahul Desai, Tyler Zhu, Samyak Parajuli, Mike Guo, Dawn Song, Jacob Steinhardt, Justin Gilmer. International Conference on Computer Vision (ICCV), 2021.

  @inproceedings{hendrycks2021many,
    author = {Dan Hendrycks and Steven Basart and Norman Mu and Saurav Kadavath and Frank Wang and Evan Dorundo and Rahul Desai and Tyler Zhu and Samyak Parajuli and Mike Guo and Dawn Song and Jacob Steinhardt and Justin Gilmer},
    booktitle = {International Conference on Computer Vision (ICCV)},
    title = {The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization},
    year = {2021},
  }
  

  [bib] [paper] [code]
  
• Natural adversarial examples. Dan Hendrycks, Kevin Zhao, Steven Basart, Jacob Steinhardt, Dawn Song. Computer Vision and Pattern Recognition (CVPR), 2021.

  @inproceedings{hendrycks2021natural,
    author = {Dan Hendrycks and Kevin Zhao and Steven Basart and Jacob Steinhardt and Dawn Song},
    booktitle = {Computer Vision and Pattern Recognition (CVPR)},
    title = {Natural adversarial examples},
    year = {2021},
  }
  

  [bib] [paper] [data]
  
• Approximating how single head attention learns. Charlie Snell, Ruiqi Zhong, Dan Klein, Jacob Steinhardt. arXiv, 2021.

  @article{snell2021approximating,
    author = {Charlie Snell and Ruiqi Zhong and Dan Klein and Jacob Steinhardt},
    journal = {arXiv},
    title = {Approximating How Single Head Attention Learns},
    year = {2021},
  }
  

  [bib] [paper]
  
• Limitations of post-hoc feature alignment for robustness. Collin Burns, Jacob Steinhardt. Computer Vision and Pattern Recognition (CVPR), 2021.

  @inproceedings{burns2021limitations,
    author = {Collin Burns and Jacob Steinhardt},
    booktitle = {Computer Vision and Pattern Recognition (CVPR)},
    title = {Limitations of Post-Hoc Feature Alignment for Robustness},
    year = {2021},
  }
  

  [bib] [paper]
  
• Measuring mathematical problem solving with the MATH dataset. Dan Hendrycks, Collin Burns, Saurav Kadavath, Akul Arora, Steven Basart, Eric Tang, Dawn Song, Jacob Steinhardt. Advances in Neural Information Processing Systems (Datasets and Benchmarks Track), 2021.

  @inproceedings{hendrycks2021measuring,
    author = {Dan Hendrycks and Collin Burns and Saurav Kadavath and Akul Arora and Steven Basart and Eric Tang and Dawn Song and Jacob Steinhardt},
    booktitle = {Advances in Neural Information Processing Systems (Datasets and Benchmarks Track)},
    title = {Measuring Mathematical Problem Solving With the {MATH} Dataset},
    year = {2021},
  }
  

  [bib] [paper] [code]
  
• Understanding generalization in adversarial training via the bias-variance decomposition. Yaodong Yu, Zitong Yang, Edgar Dobriban, Jacob Steinhardt, Yi Ma. arXiv, 2021.

  @article{yu2021understanding,
    author = {Yaodong Yu and Zitong Yang and Edgar Dobriban and Jacob Steinhardt and Yi Ma},
    journal = {arXiv},
    title = {Understanding Generalization in Adversarial Training via the Bias-Variance Decomposition},
    year = {2021},
  }
  

  [bib] [paper]
  

2020

• Enabling certification of verification-agnostic networks via memory-efficient semidefinite programming. Sumanth Dathathri, Krishnamurthy Dvijotham, Alexey Kurakin, Aditi Raghunathan, Jonathan Uesato, Rudy Bunel, Shreya Shankar, Jacob Steinhardt, Ian Goodfellow, Percy Liang, Pushmeet Kohli. Advances in Neural Information Processing Systems (NeurIPS), 2020.

  @inproceedings{dathathri2020enabling,
    author = {Sumanth Dathathri and Krishnamurthy Dvijotham and Alexey Kurakin and Aditi Raghunathan and Jonathan Uesato and Rudy Bunel and Shreya Shankar and Jacob Steinhardt and Ian Goodfellow and Percy Liang and Pushmeet Kohli},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Enabling certification of verification-agnostic networks via memory-efficient semidefinite programming},
    year = {2020},
  }
  

  [bib] [paper] [blog post] [CodaLab]
  
• How data science can ease the COVID-19 pandemic. Nigam Shah, Jacob Steinhardt. TechStream, 2020.

  @article{shah2020brookings,
    author = {Nigam Shah and Jacob Steinhardt},
    journal = {TechStream},
    title = {How data science can ease the {COVID}-19 pandemic},
    year = {2020},
  }
  

  [bib] [paper]
  
• Why robustness is key to deploying AI. Jacob Steinhardt, Helen Toner. TechStream, 2020.

  @article{steinhardt2020brookings,
    author = {Jacob Steinhardt and Helen Toner},
    journal = {TechStream},
    title = {Why robustness is key to deploying {AI}},
    year = {2020},
  }
  

  [bib] [paper]
  
• Prevalence tracking mechanisms for SARS-CoV-2. Jacob Steinhardt, Andrew Ilyas. preprint, 2020.

  @article{steinhardt2020prevalence,
    author = {Jacob Steinhardt and Andrew Ilyas},
    journal = {preprint},
    title = {Prevalence Tracking Mechanisms for {SARS}-{CoV}-2},
    year = {2020},
  }
  

  [bib] [paper]
  
• Estimation of SARS-CoV-2 infection prevalence in Santa Clara County. Steve Yadlowsky, Nigam Shah, Jacob Steinhardt. preprint, 2020.

  @article{yadlowsky2020estimation,
    author = {Steve Yadlowsky and Nigam Shah and Jacob Steinhardt},
    journal = {preprint},
    title = {Estimation of {SARS}-{CoV}-2 Infection Prevalence in {Santa} {Clara} {County}},
    year = {2020},
  }
  

  [bib] [paper] [code] [blog post]
  
• Identifying statistical bias in dataset replication. Logan Engstrom, Andrew Ilyas, Shibani Santurkar, Dimitris Tsipras, Jacob Steinhardt, Aleksander Madry. International Conference on Machine Learning (ICML), 2020.

  @inproceedings{engstrom2020identifying,
    author = {Logan Engstrom and Andrew Ilyas and Shibani Santurkar and Dimitris Tsipras and Jacob Steinhardt and Aleksander Madry},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Identifying Statistical Bias in Dataset Replication},
    year = {2020},
  }
  

  [bib] [paper] [code] [blog post]
  
• Rethinking bias-variance trade-off for generalization of neural networks. Zitong Yang, Yaodong Yu, Chong You, Jacob Steinhardt, Yi Ma. International Conference on Machine Learning (ICML), 2020.

  @inproceedings{yang2020rethinking,
    author = {Zitong Yang and Yaodong Yu and Chong You and Jacob Steinhardt and Yi Ma},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Rethinking bias-variance trade-off for generalization of neural networks},
    year = {2020},
  }
  

  [bib] [paper]
  
• When does the tukey median work? Banghua Zhu, Jiantao Jiao, Jacob Steinhardt. IEEE International Symposium on Information Theory (ISIT), 2020.

  @inproceedings{zhu2020tukey,
    author = {Banghua Zhu and Jiantao Jiao and Jacob Steinhardt},
    booktitle = {IEEE International Symposium on Information Theory (ISIT)},
    title = {When does the Tukey median work?},
    year = {2020},
  }
  

  [bib] [paper]
  

2019

• Testing robustness against unforeseen adversaries. Daniel Kang, Yi Sun, Dan Hendrycks, Tom Brown, Jacob Steinhardt. arXiv, 2019.

  @article{kang2019testing,
    author = {Daniel Kang and Yi Sun and Dan Hendrycks and Tom Brown and Jacob Steinhardt},
    journal = {arXiv},
    title = {Testing Robustness Against Unforeseen Adversaries},
    year = {2019},
  }
  

  [bib] [paper] [reviews] [code]
  
• Sever: a robust meta-algorithm for stochastic optimization. Ilias Diakonikolas, Gautam Kamath, Daniel M. Kane, Jerry Li, Jacob Steinhardt, Alistair Stewart. International Conference on Machine Learning (ICML), 2019.

  @inproceedings{diakonikolas2019sever,
    author = {Ilias Diakonikolas and Gautam Kamath and Daniel M. Kane and Jerry Li and Jacob Steinhardt and Alistair Stewart},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Sever: A Robust Meta-Algorithm for Stochastic Optimization},
    year = {2019},
  }
  

  [bib] [paper] [code]
  
• Troubling trends in machine learning scholarship. Zachary C. Lipton, Jacob Steinhardt. ACM Queue, 2019.

  @article{lipton2019troubling,
    author = {Zachary C. Lipton and Jacob Steinhardt},
    journal = {ACM Queue},
    title = {Troubling Trends in Machine Learning Scholarship},
    volume = {17},
    year = {2019},
  }
  

  [bib] [paper]
  
• FrAngel: component-based synthesis with control structures. Kensen Shi, Jacob Steinhardt, Percy Liang. Principles of Programming Languages (POPL), 2019.

  @inproceedings{shi2019frangel,
    author = {Kensen Shi and Jacob Steinhardt and Percy Liang},
    booktitle = {Principles of Programming Languages (POPL)},
    title = \{\{F}r{A}ngel: Component-Based Synthesis with Control Structures},
    year = {2019},
  }
  

  [bib] [paper] [CodaLab]
  

2018

• Better agnostic clustering via relaxed tensor norms. Pravesh Kothari, Jacob Steinhardt. Symposium on Theory of Computing (STOC), 2018.

  @inproceedings{kothari2018agnostic,
    author = {Pravesh Kothari and Jacob Steinhardt},
    booktitle = {Symposium on Theory of Computing (STOC)},
    title = {Better Agnostic Clustering via Relaxed Tensor Norms},
    year = {2018},
  }
  

  [bib] [paper]
  
• Resilience: a criterion for learning in the presence of arbitrary outliers. Jacob Steinhardt, Moses Charikar, Gregory Valiant. Innovations in Theoretical Computer Science (ITCS), 2018.

  @inproceedings{steinhardt2018resilience,
    author = {Jacob Steinhardt and Moses Charikar and Gregory Valiant},
    booktitle = {Innovations in Theoretical Computer Science (ITCS)},
    title = {Resilience: A Criterion for Learning in the Presence of Arbitrary Outliers},
    year = {2018},
  }
  

  [bib] [paper] [talk] [slides]
  
• Robust learning: information theory and algorithms. Jacob Steinhardt. Stanford University, 2018.

  @phdthesis{steinhardt2018thesis,
    author = {Jacob Steinhardt},
    school = {Stanford University},
    title = {Robust Learning: Information Theory and Algorithms},
    year = {2018},
  }
  

  [bib] [thesis]
  
• The malicious use of artificial intelligence: forecasting, prevention, and mitigation. Miles Brundage, Shahar Avin, Jack Clark, Helen Toner, Peter Eckersley, Ben Garfinkel, Allan Dafoe, Paul Scharre, Thomas Zeitzoff, Bobby Filar, others. arXiv, 2018.

  @article{brundage2018malicious,
    author = {Miles Brundage and Shahar Avin and Jack Clark and Helen Toner and Peter Eckersley and Ben Garfinkel and Allan Dafoe and Paul Scharre and Thomas Zeitzoff and Bobby Filar and others},
    journal = {arXiv},
    title = {The malicious use of artificial intelligence: Forecasting, prevention, and mitigation},
    year = {2018},
  }
  

  [bib] [paper]
  
• Semidefinite relaxations for certifying robustness to adversarial examples. Aditi Raghunathan, Jacob Steinhardt, Percy Liang. Advances in Neural Information Processing Systems (NeurIPS), 2018.

  @inproceedings{raghunathan2018sdp,
    author = {Aditi Raghunathan and Jacob Steinhardt and Percy Liang},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Semidefinite relaxations for certifying robustness to adversarial examples},
    year = {2018},
  }
  

  [bib] [paper]
  
• Certified defenses against adversarial examples. Aditi Raghunathan, Jacob Steinhardt, Percy Liang. International Conference on Learning Representations (ICLR), 2018.

  @inproceedings{raghunathan2018certified,
    author = {Aditi Raghunathan and Jacob Steinhardt and Percy Liang},
    booktitle = {International Conference on Learning Representations (ICLR)},
    title = {Certified defenses against adversarial examples},
    year = {2018},
  }
  

  [bib] [paper] [reviews] [CodaLab]
  

2017

• Learning from untrusted data. Moses Charikar, Jacob Steinhardt, Gregory Valiant. Symposium on Theory of Computing (STOC), 2017.

  @inproceedings{charikar2017learning,
    author = {Moses Charikar and Jacob Steinhardt and Gregory Valiant},
    booktitle = {Symposium on Theory of Computing (STOC)},
    title = {Learning from Untrusted Data},
    year = {2017},
  }
  

  [bib] [paper] [talk] [slides] [poster]
  
• Does robustness imply tractability? A lower bound for planted clique in the semi-random model. Jacob Steinhardt. arXiv, 2017.

  @article{steinhardt2017clique,
    author = {Jacob Steinhardt},
    journal = {arXiv},
    title = {Does robustness imply tractability? {A} lower bound for planted clique in the semi-random model},
    year = {2017},
  }
  

  [bib] [paper]
  
• Certified defenses for data poisoning attacks. Jacob Steinhardt, Pang Wei Koh, Percy Liang. Advances in Neural Information Processing Systems (NeurIPS), 2017.

  @inproceedings{steinhardt2017certified,
    author = {Jacob Steinhardt and Pang Wei Koh and Percy Liang},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Certified Defenses for Data Poisoning Attacks},
    year = {2017},
  }
  

  [bib] [paper] [code] [CodaLab]
  

2016

• Memory, communication, and statistical queries. Jacob Steinhardt, Gregory Valiant, Stefan Wager. Conference on Learning Theory (COLT), 2016.

  @inproceedings{steinhardt2016memory,
    author = {Jacob Steinhardt and Gregory Valiant and Stefan Wager},
    booktitle = {Conference on Learning Theory (COLT)},
    title = {Memory, Communication, and Statistical Queries},
    year = {2016},
  }
  

  [bib] [paper]
  
• Avoiding imposters and delinquents: adversarial crowdsourcing and peer prediction. Jacob Steinhardt, Gregory Valiant, Moses Charikar. Advances in Neural Information Processing Systems (NeurIPS), 2016.

  @inproceedings{steinhardt2016avoiding,
    author = {Jacob Steinhardt and Gregory Valiant and Moses Charikar},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Avoiding Imposters and Delinquents: Adversarial Crowdsourcing and Peer Prediction},
    year = {2016},
  }
  

  [bib] [paper]
  
• Concrete problems in AI safety. Dario Amodei, Chris Olah, Jacob Steinhardt, Paul Christiano, John Schulman, Dan Mané. arXiv, 2016.

  @article{amodei2016concrete,
    author = {Dario Amodei and Chris Olah and Jacob Steinhardt and Paul Christiano and John Schulman and Dan Mané},
    journal = {arXiv},
    title = {Concrete problems in {AI} safety},
    year = {2016},
  }
  

  [bib] [paper]
  
• Unsupervised risk estimation using only conditional independence structure. Jacob Steinhardt, Percy Liang. Advances in Neural Information Processing Systems (NeurIPS), 2016.

  @inproceedings{steinhardt2016risk,
    author = {Jacob Steinhardt and Percy Liang},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Unsupervised Risk Estimation Using Only Conditional Independence Structure},
    year = {2016},
  }
  

  [bib] [paper]
  

2015

• Minimax rates for memory-bounded sparse linear regression. Jacob Steinhardt, John Duchi. Conference on Learning Theory (COLT), 2015.

  @inproceedings{steinhardt2015minimax,
    author = {Jacob Steinhardt and John Duchi},
    booktitle = {Conference on Learning Theory (COLT)},
    title = {Minimax rates for memory-bounded sparse linear regression},
    year = {2015},
  }
  

  [bib] [paper] [talk] [slides] [poster]
  
• Learning with relaxed supervision. Jacob Steinhardt, Percy Liang. Advances in Neural Information Processing Systems (NeurIPS), 2015.

  @inproceedings{steinhardt2015relaxed,
    author = {Jacob Steinhardt and Percy Liang},
    booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
    title = {Learning with Relaxed Supervision},
    year = {2015},
  }
  

  [bib] [paper] [poster] [CodaLab]
  
• Reified context models. Jacob Steinhardt, Percy Liang. International Conference on Machine Learning (ICML), 2015.

  @inproceedings{steinhardt2015rcm,
    author = {Jacob Steinhardt and Percy Liang},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Reified Context Models},
    year = {2015},
  }
  

  [bib] [paper] [talk] [slides] [poster] [CodaLab]
  
• Learning fast-mixing models for structured prediction. Jacob Steinhardt, Percy Liang. International Conference on Machine Learning (ICML), 2015.

  @inproceedings{steinhardt2015fast,
    author = {Jacob Steinhardt and Percy Liang},
    booktitle = {International Conference on Machine Learning (ICML)},
    pages = {1063--1072},
    title = {Learning Fast-Mixing Models for Structured Prediction},
    year = {2015},
  }
  

  [bib] [paper] [talk] [slides] [poster] [CodaLab]
  
• Learning where to sample in structured prediction. Tianlin Shi, Jacob Steinhardt, Percy Liang. Artificial Intelligence and Statistics (AISTATS), 2015.

  @inproceedings{shi2015sample,
    author = {Tianlin Shi and Jacob Steinhardt and Percy Liang},
    booktitle = {Artificial Intelligence and Statistics (AISTATS)},
    pages = {875--884},
    title = {Learning Where To Sample in Structured Prediction},
    year = {2015},
  }
  

  [bib] [paper] [slides] [CodaLab]
  

2014

• The statistics of streaming sparse regression. Jacob Steinhardt, Stefan Wager, Percy Liang. arXiv preprint arXiv:1412.4182, 2014.

  @article{steinhardt2014sparse,
    author = {Jacob Steinhardt and Stefan Wager and Percy Liang},
    journal = {arXiv preprint arXiv:1412.4182},
    title = {The Statistics of Streaming Sparse Regression},
    year = {2014},
  }
  

  [bib] [paper]
  
• Adaptivity and optimism: an improved exponentiated gradient algorithm. Jacob Steinhardt, Percy Liang. International Conference on Machine Learning (ICML), 2014.

  @inproceedings{steinhardt2014eg,
    author = {Jacob Steinhardt and Percy Liang},
    booktitle = {International Conference on Machine Learning (ICML)},
    title = {Adaptivity and Optimism: An Improved Exponentiated Gradient Algorithm},
    year = {2014},
  }
  

  [bib] [paper] [slides] [poster]
  
• Filtering with abstract particles. Jacob Steinhardt, Percy Liang. International Conference on Machine Learning (ICML), 2014.

  @inproceedings{steinhardt2014filtering,
    author = {Jacob Steinhardt and Percy Liang},
    booktitle = {International Conference on Machine Learning (ICML)},
    pages = {727--735},
    title = {Filtering with Abstract Particles},
    year = {2014},
  }
  

  [bib] [paper] [supplemental material] [slides] [poster]
  

2012

• Flexible martingale priors for deep hierarchies. Jacob Steinhardt, Zoubin Ghahramani. Artificial Intelligence and Statistics (AISTATS), 2012.

  @inproceedings{steinhardt2012flexible,
    author = {Jacob Steinhardt and Zoubin Ghahramani},
    booktitle = {Artificial Intelligence and Statistics (AISTATS)},
    title = {Flexible martingale priors for deep hierarchies},
    year = {2012},
  }
  

  [bib] [paper] [slides] [poster]
  

2011

• Finite-time regional verification of stochastic nonlinear systems. Jacob Steinhardt, Russ Tedrake. Robotics: Science and Systems (RSS), 2011.

  @inproceedings{steinhardt2011finite,
    author = {Jacob Steinhardt and Russ Tedrake},
    booktitle = {Robotics: Science and Systems (RSS)},
    title = {Finite-time regional verification of stochastic nonlinear systems},
    year = {2011},
  }
  

  [bib] [paper] [journal] [slides] [poster]
  
• Pathological properties of deep bayesian hierarchies. Jacob Steinhardt, Zoubin Ghahramani. NIPS Workshop on Bayesian Nonparametrics, 2011.

  @inproceedings{steinhardt2011pathological,
    author = {Jacob Steinhardt and Zoubin Ghahramani},
    booktitle = {NIPS Workshop on Bayesian Nonparametrics},
    title = {Pathological Properties of Deep Bayesian Hierarchies},
    year = {2011},
  }
  

  [bib] [paper] [poster]
  

2010

• Permutations with ascending and descending blocks. Jacob Steinhardt. Electronic Journal of Combinatorics, 2010.

  @article{steinhardt2010permutations,
    author = {Jacob Steinhardt},
    journal = {Electronic Journal of Combinatorics},
    title = {Permutations with ascending and descending blocks},
    volume = {17},
    year = {2010},
  }
  

  [bib] [paper] [slides]
  

2009

• On coloring the odd-distance graph. Jacob Steinhardt. Electronic Journal of Combinatorics, 2009.

  @article{steinhardt2009coloring,
    author = {Jacob Steinhardt},
    journal = {Electronic Journal of Combinatorics},
    title = {On coloring the odd-distance graph},
    volume = {16},
    year = {2009},
  }
  

  [bib] [paper]
  

2007

• Cayley graphs formed by conjugate generating sets of s_n. Jacob Steinhardt. preprint, 2007.

  @article{steinhardt2007cayley,
    author = {Jacob Steinhardt},
    journal = {preprint},
    title = {Cayley Graphs Formed by Conjugate Generating Sets of S_n},
    year = {2007},
  }
  

  [bib] [paper]