Scholar
Chris Cundy
Google Scholar ID: TWAn5XoAAAAJ
FAR.AI
Citations & Impact
Citations
1,318
H-index
11
i10-index
13
Publications
20
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- - Publication: Preference Learning with Lie Detectors can Induce Honesty or Evasion
- - Publication: SequenceMatch: Imitation Learning for Autoregressive Sequence Modelling with Backtracking
- - Publication: Privacy-Constrained Policies via Mutual Information Regularized Policy Gradients
- - Publication: LMPriors: Pre-Trained Language Models as Task-Specific Priors
- - Publication: IQ-Learn: Inverse soft-Q Learning for Imitation
Research Experience
- - Research Scientist at FAR AI
- - During his PhD, he studied a diverse range of topics including constrained reinforcement learning, variational inference, and autoregressive models
- - Interned at the Centre for Human Compatible AI, the Future of Humanity Institute at Oxford University, and DeepMind
Education
- - PhD in Computer Science, 2018-2024, Stanford University, advised by Stefano Ermon
- - MEng in Computer Science, 2017, Cambridge University
- - BA in Natural Sciences (Physics), 2016, Cambridge University, worked under Carl E. Rasmussen
Background
- A Research Scientist focusing on reducing catastrophic risks from advanced AI systems. Interests include Deceptive Behavior from LLMs, Risk Evaluation and Elicitation, Governance of Frontier Models, Adversarial Robustness, and Probabilistic Machine Learning.
Miscellany
- Contact via email: chris dot j dot cundy at gmail dot com.
Co-authors