Arvind Narayanan

Learn best practices for developing AI models responsibly and effectively.

Investigating how AI agents reproduce scientific results through a new benchmark.

Investigating the limits of repeated sampling in weaker language models.