Alexander Pan

I work on AI safety at Meta. Previously, I led safety at xAI.

I did my PhD at Berkeley, where I was lucky to have Jacob Steinhardt as my advisor. Before that, I studied math and CS at Caltech.

Selected Publications

LatentQA: Teaching LLMs to Decode Activations Into Natural Language

Alexander Pan, Lijie Chen, Jacob Steinhardt

ICLR 2026·pdf / code

The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning

Nathaniel Li*, Alexander Pan*, …, Alexandr Wang**, Dan Hendrycks**

ICML 2024·pdf / code

Feedback Loops With Language Models Drive In-Context Reward Hacking

Alexander Pan, Erik Jones, Meena Jagadeesan, Jacob Steinhardt

ICML 2024·pdf / code

Do the Rewards Justify the Means? Measuring Trade-Offs Between Rewards and Ethical Behavior in the MACHIAVELLI Benchmark

Alexander Pan*, Jun Shern Chan*, Andy Zou*, …, Dan Hendrycks

ICML 2023 Oral·pdf / code

The Effects of Reward Misspecification: Mapping and Mitigating Misaligned Models

Alexander Pan, Kush Bhatia, Jacob Steinhardt

ICLR 2022·pdf / code

Contact

Feel free to email me at aypan (dot) 17 (at) berkeley (dot) edu.