Elamur Research⌕
☰
home / authors / Jacob Steinhardt / all works

Jacob Steinhardt — all works

128 works · page 1 of 2 · 1–100 · ← Jacob Steinhardt

1 2

  1. Measuring Massive Multitask Language UnderstandingarXiv:2009.03300Dan Hendrycks, Collin Burns, Steven Basart, et al.2020 · 3,620 citations
  2. Measuring Mathematical Problem Solving With the MATH DatasetarXiv:2103.03874Dan Hendrycks, Collin Burns, Saurav Kadavath, et al.2021 · 2,395 citations
  3. The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization ↗IEEE/CVF International Conference on Computer Vision (ICCV) · no summary yetDan Hendrycks et al.2021 · 1,291 citations
  4. Natural Adversarial Examples ↗IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · no summary yetDan Hendrycks et al.2019 · 1,121 citations
  5. The Malicious Use of Artificial Intelligence: Forecasting, Prevention, and Mitigation ↗arXiv (Cornell University) · no summary yetMiles Brundage et al.2018 · 866 citations
  6. Jailbroken: How Does LLM Safety Training Fail?Conference on Neural Information Processing Systems (NeurIPS)Alexander Wei, Nika Haghtalab, Jacob Steinhardt2023 · 688 citations
  7. Certified Defenses against Adversarial Examples ↗arXiv (Cornell University) · no summary yetAditi Raghunathan, Jacob Steinhardt, Percy Liang2018 · 550 citations
  8. Measuring Coding Challenge Competence With APPSarXiv:2105.09938Dan Hendrycks, Steven Basart, Saurav Kadavath, et al.2021 · 457 citations
  9. Aligning AI With Shared Human ValuesarXiv:2008.02275Dan Hendrycks, Collin Burns, Steven Basart, et al.2020 · 395 citations
  10. Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small ↗arXiv:2211.00593 · no summary yetKevin Wang et al.2022 · 376 citations
  11. Troubling Trends in Machine Learning Scholarship ↗arXiv:1807.03341 · no summary yetZachary C. Lipton, Jacob Steinhardt2018 · 257 citations
  12. Discovering Latent Knowledge in Language Models Without Supervision ↗arXiv:2212.03827 · no summary yetCollin Burns et al.2022 · 250 citations
  13. Concrete Problems in AI SafetyarXiv:1606.06565Dario Amodei, Chris Olah, Jacob Steinhardt, et al.2016 · 232 citations
  14. Progress measures for grokking via mechanistic interpretabilityarXiv:2301.05217Neel Nanda, Lawrence Chan, Tom Lieberum, et al.2023 · 228 citations
  15. Learning from untrusted data ↗ACM SIGACT Symposium on Theory of Computing (STOC) · no summary yetMoses Charikar, Jacob Steinhardt, Gregory Valiant2017 · 218 citations
  16. Semidefinite relaxations for certifying robustness to adversarial examples ↗arXiv (Cornell University) · no summary yetAditi Raghunathan, Jacob Steinhardt, Percy Liang2018 · 201 citations
  17. Stronger data poisoning attacks break data sanitization defenses ↗Machine Learning · no summary yetPang Wei Koh, Jacob Steinhardt, Percy Liang2018 · 200 citations
  18. Eliciting Latent Predictions from Transformers with the Tuned LensarXiv:2303.08112Nora Belrose, Igor Ostrovsky, Lev McKinney, et al.2023 · 188 citations
  19. Scaling Out-of-Distribution Detection for Real-World Settings ↗arXiv (Cornell University) · no summary yetDan Hendrycks et al.2019 · 175 citations
  20. Unsolved Problems in ML Safety ↗arXiv:2109.13916 · no summary yetDan Hendrycks et al.2021 · 151 citations
  21. Sever: A Robust Meta-Algorithm for Stochastic Optimization ↗arXiv (Cornell University) · no summary yetIlias Diakonikolas et al.2018 · 139 citations
  22. PixMix: Dreamlike Pictures Comprehensively Improve Safety Measures ↗IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · no summary yetDan Hendrycks et al.2022 · 111 citations
  23. Certified Defenses for Data Poisoning Attacks ↗arXiv (Cornell University) · no summary yetJacob Steinhardt, Pang Wei Koh, Percy Liang2017 · 103 citations
  24. Robust moment estimation and improved clustering via sum of squares ↗ACM SIGACT Symposium on Theory of Computing (STOC) · no summary yetPravesh K. Kothari, Jacob Steinhardt, David Steurer2018 · 92 citations
  25. Automatically Auditing Large Language Models via Discrete Optimization ↗arXiv:2303.04381 · no summary yetErik Jones et al.2023 · 89 citations
  26. Finite-time regional verification of stochastic non-linear systems ↗The International Journal of Robotics Research · no summary yetJacob Steinhardt, Russ Tedrake2012 · 89 citations
  27. The Effects of Reward Misspecification: Mapping and Mitigating Misaligned Models ↗arXiv (Cornell University) · no summary yetPan, Alexander, Kush Bhatia, Jacob Steinhardt2022 · 83 citations
  28. Capturing Failures of Large Language Models Via Human Cognitive Biases ↗neural information processing systems · no summary yetErik Jones, Jacob Steinhardt2022 · 82 citations
  29. Rethinking Bias-Variance Trade-off for Generalization of Neural Networks ↗arXiv (Cornell University) · no summary yetZitong Yang et al.2020 · 81 citations
  30. Testing Robustness Against Unforeseen Adversaries ↗arXiv (Cornell University) · no summary yetKaufmann, Max et al.2019 · 80 citations
  31. Resilience: A Criterion for Learning in the Presence of Arbitrary Outliers ↗arXiv (Cornell University) · no summary yetJacob Steinhardt, Moses Charikar, Gregory Valiant2017 · 73 citations
  32. Protein language models are biased by unequal sequence sampling across the tree of life ↗bioRxiv (Cold Spring Harbor Laboratory) · no summary yetFrances Ding, Jacob Steinhardt2024 · 57 citations
  33. Adaptivity and Optimism: An Improved Exponentiated Gradient Algorithm ↗no summary yetJacob Steinhardt, Percy Liang2014 · 48 citations
  34. FrAngel: component-based synthesis with control structures ↗Proceedings of the ACM on Programming Languages · no summary yetKensen Shi, Jacob Steinhardt, Percy Liang2019 · 44 citations
  35. Overthinking the Truth: Understanding how Language Models Process False Demonstrations ↗arXiv (Cornell University) · no summary yetDanny Halawi, Jean-Stanislas Denain, Jacob Steinhardt2023 · 43 citations
  36. Goal Driven Discovery of Distributional Differences via Language DescriptionsarXiv:2302.14233Ruiqi Zhong, Peter Zhang, Steve Li, et al.2023 · 42 citations
  37. Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations ↗arXiv (Cornell University) · no summary yetYanda Chen et al.2023 · 41 citations
  38. Are Larger Pretrained Language Models Uniformly Better? Comparing Performance at the Instance Level ↗Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021 · no summary yetRuiqi Zhong et al.2021 · 39 citations
  39. Approaching Human-Level Forecasting with Language Models ↗neural information processing systems · no summary yetDanny Halawi et al.2024 · 38 citations
  40. Describing Differences between Text Distributions with Natural Language ↗arXiv (Cornell University) · no summary yetRuiqi Zhong et al.2022 · 34 citations
  41. Generalized resilience and robust statistics ↗The Annals of Statistics · no summary yetBanghua Zhu, Jiantao Jiao, Jacob Steinhardt2019 · 34 citations
  42. Language Models Learn to Mislead Humans via RLHF ↗arXiv (Cornell University) · no summary yetJiaxin Wen et al.2024 · 33 citations
  43. How do Language Models Bind Entities in Context? ↗arXiv (Cornell University) · no summary yetJiahai Feng, Jacob Steinhardt2023 · 31 citations
  44. What Would Jiminy Cricket Do? Towards Agents That Behave Morally ↗arXiv (Cornell University) · no summary yetDan Hendrycks et al.2021 · 30 citations
  45. Describing Differences in Image Sets with Natural Language ↗IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · no summary yetLisa Dunlap et al.2024 · 28 citations
  46. Feedback Loops With Language Models Drive In-Context Reward Hacking ↗arXiv:2402.06627 · no summary yetAlexander Pan et al.2024 · 26 citations
  47. Better Agnostic Clustering Via Relaxed Tensor Norms ↗arXiv (Cornell University) · no summary yetPravesh K. Kothari, Jacob Steinhardt2017 · 25 citations
  48. Mass-Producing Failures of Multimodal Systems with Language Models ↗neural information processing systems · no summary yetShengbang Tong, Erik Jones, Jacob Steinhardt2023 · 23 citations
  49. Interpreting CLIP's Image Representation via Text-Based Decomposition ↗arXiv (Cornell University) · no summary yetYossi Gandelsman, Alexei A. Efros, Jacob Steinhardt2023 · 22 citations
  50. Robust estimation via generalized quasi-gradients ↗Information and Inference A Journal of the IMA · no summary yetBanghua Zhu, Jiantao Jiao, Jacob Steinhardt2021 · 22 citations
  51. Transfer of Adversarial Robustness Between Perturbation Types ↗arXiv (Cornell University) · no summary yetDaniel D. Kang et al.2019 · 22 citations
  52. Research for practice ↗Communications of the ACM · no summary yetZachary C. Lipton, Jacob Steinhardt2019 · 22 citations
  53. Enabling certification of verification-agnostic networks via memory-efficient semidefinite programming ↗arXiv:2010.11645 · no summary yetSumanth Dathathri et al.2020 · 21 citations
  54. Minimax rates for memory-bounded sparse linear regression ↗Conference on Learning Theory · no summary yetJacob Steinhardt, John C. Duchi2015 · 21 citations
  55. Unsupervised Risk Estimation Using Only Conditional Independence Structure ↗arXiv (Cornell University) · no summary yetJacob Steinhardt, Percy Liang2016 · 20 citations
  56. Forecasting Future World Events with Neural Networks ↗neural information processing systems · no summary yetAndy Zou et al.2022 · 18 citations
  57. Covert Malicious Finetuning: Challenges in Safeguarding LLM Adaptation ↗arXiv (Cornell University) · no summary yetDanny Halawi et al.2024 · 17 citations
  58. When does the Tukey Median work? ↗IEEE International Symposium on Information Theory (ISIT) · no summary yetBanghua Zhu, Jiantao Jiao, Jacob Steinhardt2020 · 16 citations
  59. Learning Where to Sample in Structured Prediction ↗no summary yetTianlin Shi, Jacob Steinhardt, Percy Liang2015 · 15 citations
  60. Learning with relaxed supervision ↗neural information processing systems · no summary yetJacob Steinhardt, Percy Liang2015 · 15 citations
  61. Identifying Statistical Bias in Dataset Replication ↗arXiv (Cornell University) · no summary yetLogan Engstrom et al.2020 · 14 citations
  62. Permutations with Ascending and Descending Blocks ↗The Electronic Journal of Combinatorics · no summary yetJacob Steinhardt2010 · 14 citations
  63. Which Attention Heads Matter for In-Context Learning? ↗arXiv (Cornell University) · no summary yetKayo Yin, Jacob Steinhardt2025 · 12 citations
  64. More Than a Toy: Random Matrix Models Predict How Real-World Neural Representations Generalize ↗arXiv (Cornell University) · no summary yetWei, Alexander, Wei Ping Hu, Jacob Steinhardt2022 · 12 citations
  65. Limitations of Post-Hoc Feature Alignment for Robustness ↗IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · no summary yetCollin Burns, Jacob Steinhardt2021 · 12 citations
  66. LatentQA: Teaching LLMs to Decode Activations Into Natural Language ↗arXiv (Cornell University) · no summary yetPan, Alexander, Lijie Chen, Jacob Steinhardt2024 · 11 citations
  67. Article-Level Slant and Polarization of News Consumption on Social Media ↗SSRN Electronic Journal · no summary yetLuca Braghieri et al.2024 · 11 citations
  68. Supply-Side Equilibria in Recommender Systems ↗neural information processing systems · no summary yetMeena Jagadeesan, Nikhil Garg, Jacob Steinhardt2023 · 11 citations
  69. Understanding Generalization in Adversarial Training via the Bias-Variance Decomposition ↗arXiv (Cornell University) · no summary yetYaodong Yu et al.2021 · 11 citations
  70. On Coloring the Odd-Distance Graph ↗The Electronic Journal of Combinatorics · no summary yetJacob Steinhardt2009 · 11 citations
  71. Monitoring Latent World States in Language Models with Propositional Probes ↗arXiv (Cornell University) · no summary yetJiahai Feng, Stuart Russell, Jacob Steinhardt2024 · 10 citations
  72. Does robustness imply tractability? A lower bound for planted clique in the semi-random model ↗arXiv (Cornell University) · no summary yetJacob Steinhardt2017 · 10 citations
  73. VibeCheck: Discover and Quantify Qualitative Differences in Large Language Models ↗arXiv (Cornell University) · no summary yetLisa Dunlap et al.2024 · 9 citations
  74. Explaining Datasets in Words: Statistical Models with Natural Language Parameters ↗neural information processing systems · no summary yetRuiqi Zhong et al.2024 · 9 citations
  75. Grounding Representation Similarity Through Statistical Testing ↗arXiv (Cornell University) · no summary yetFrances Ding, Jean-Stanislas Denain, Jacob Steinhardt2021 · 9 citations
  76. Filtering with Abstract Particles ↗no summary yetJacob Steinhardt, Percy Liang2014 · 9 citations
  77. Eliciting Language Model Behaviors with Investigator Agents ↗arXiv (Cornell University) · no summary yetXiang Lisa Li et al.2025 · 8 citations
  78. Predicting Out-of-Distribution Error with the Projection Norm ↗arXiv (Cornell University) · no summary yetYaodong Yu et al.2022 · 8 citations
  79. Memory, Communication, and Statistical Queries ↗Electronic colloquium on computational complexity · no summary yetJacob Steinhardt, Gregory Valiant, Stefan Wager2015 · 8 citations
  80. Flexible Martingale Priors for Deep Hierarchies. ↗no summary yetJacob Steinhardt, Zoubin Ghahramani2012 · 8 citations
  81. Finite-Time Regional Verification of Stochastic Nonlinear Systems ↗no summary yetJacob Steinhardt, Russ Tedrake2011 · 8 citations
  82. Training Language Models to Explain Their Own Computations ↗arXiv (Cornell University) · no summary yetBelinda Z. Li et al.2025 · 7 citations
  83. Establishing Best Practices for Building Rigorous Agentic Benchmarks ↗arXiv:2507.02825 · no summary yetYuxuan Zhu et al.2025 · 7 citations
  84. Avoiding Imposters and Delinquents: Adversarial Crowdsourcing and Peer Prediction ↗arXiv (Cornell University) · no summary yetJacob Steinhardt, Gregory Valiant, Moses Charikar2016 · 7 citations
  85. Reified Context Models ↗arXiv (Cornell University) · no summary yetJacob Steinhardt, Percy Liang2015 · 6 citations
  86. The Statistics of Streaming Sparse Regression ↗arXiv (Cornell University) · no summary yetJacob Steinhardt, Stefan Wager, Percy Liang2014 · 6 citations
  87. Predictive Concept Decoders: Training Scalable End-to-End Interpretability Assistants ↗arXiv (Cornell University) · no summary yetHuang, Vincent et al.2025 · 5 citations
  88. Adversaries Can Misuse Combinations of Safe Models ↗arXiv (Cornell University) · no summary yetErik Jones, Anca D. Dragan, Jacob Steinhardt2024 · 5 citations
  89. Learning Equilibria in Matching Markets with Bandit Feedback ↗Journal of the ACM · no summary yetMeena Jagadeesan et al.2021 · 5 citations
  90. Approximating How Single Head Attention Learns ↗arXiv (Cornell University) · no summary yetCharlie Snell et al.2021 · 5 citations
  91. Language Model Circuits Are Sparse in the Neuron Basis ↗arXiv (Cornell University) · no summary yetAryaman Arora et al.2026 · 4 citations
  92. Understanding In-context Learning of Addition via Activation Subspaces ↗arXiv (Cornell University) · no summary yetXinyan Hu et al.2025 · 4 citations
  93. Establishing Best Practices in Building Rigorous Agentic Benchmarks ↗neural information processing systems · no summary yetYuxuan Zhu et al.2025 · 4 citations
  94. Extractive Structures Learned in Pretraining Enable Generalization on Finetuned Facts ↗arXiv (Cornell University) · no summary yetJiahai Feng, Stuart Russell, Jacob Steinhardt2024 · 4 citations
  95. What Do Learning Dynamics Reveal About Generalization in LLM Reasoning? ↗arXiv (Cornell University) · no summary yetKatie Kang et al.2024 · 4 citations
  96. Improved Bayes Risk Can Yield Reduced Social Welfare Under Competition ↗neural information processing systems · no summary yetMeena Jagadeesan et al.2023 · 4 citations
  97. How Would the Viewer Feel? Estimating Wellbeing from Video Scenarios ↗neural information processing systems · no summary yetMantas Mazeika et al.2022 · 4 citations
  98. Learning Fast-Mixing Models for Structured Prediction ↗arXiv (Cornell University) · no summary yetJacob Steinhardt, Percy Liang2015 · 4 citations
  99. Interpreting the Second-Order Effects of Neurons in CLIP ↗arXiv (Cornell University) · no summary yetYossi Gandelsman, Alexei A. Efros, Jacob Steinhardt2024 · 3 citations
  100. Auditing Visualizations: Transparency Methods Struggle to Detect Anomalous Behavior ↗arXiv (Cornell University) · no summary yetJean-Stanislas Denain, Jacob Steinhardt2022 · 3 citations

1 2

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/authors/jacob-steinhardt/works/