Caleb Biddulph

Message

MATS 8.1 scholar, mentored by Micah Carroll. Former SWE at Google Gemini. Founding president of Cornell Effective Altruism.

I have not signed any contracts that I can't mention exist, as of December 20, 2025. I'll try to update this statement at least once a year, so...

1701

237

5 ways to improve CoT faithfulness

In the case for CoT unfaithfulness is overstated, @nostalgebraist pointed out that reading the chain-of-thought (CoT) reasoning of models is neglected as an interpretability technique. I completely agree, but I also believe that with our current methods, there's a significant risk of being misled by the CoTs of models like...

Oct 5, 202446

Preparing for AI-assisted alignment research: we need data!

Thanks to Neel Nanda, Madeleine Chang, Michael Chen, Oliver Zhang, and Aidan O’Gara for their early feedback on this post. Summary: Alignment researchers could record the intermediate outputs of their work in anticipation of using the data to train future AI models. These models could eventually accelerate alignment research by...

Jan 17, 202331

AI ALIGNMENT FORUM
AF

AI ALIGNMENT FORUM
AF

Caleb Biddulph

Caleb Biddulph

Caleb Biddulph

Caleb Biddulph

5 ways to improve CoT faithfulness

Preparing for AI-assisted alignment research: we need data!

5 ways to improve CoT faithfulness

Preparing for AI-assisted alignment research: we need data!