Matthew Kowal is an AI interpretability researcher working on mechanistic interpretability.
How media typically covers Matthew Kowal
Matthew Kowal as author
Large language models can effectively persuade people to believe conspiracy theories as readily as debunk them, and OpenAI's guardrails do little to prevent this misuse.
“Author of "Large Language Models Can Effectively Convince People to Believe Conspiracies" on arXiv”
Referenced in coverage
“Co-author of research on interpretability-based training protocols”