Josh Engels — AI Alignment Forum

Test your best methods on our hard CoT interp tasks

Authors: Daria Ivanova, Riya Tyagi, Josh Engels, Neel Nanda Daria and Riya are co-first authors. This work was done during Neel Nanda’s MATS 9.0. Claude helped write code and suggest edits for this post. Most of our tasks fall in 3 categories: predicting future actions, detecting the effect of an...

Mar 2657

Brief Explorations in LLM Value Rankings

Code and data can be found here Executive Summary * We use data from Zhang et al. (2025) to measure LLM values. We find that our value metric can sometimes predict LLM behaviors on a test distribution in non-safety-relevant settings, but it is not super consistent. * In Zhang et...

Jan 1239

Steering RL Training: Benchmarking Interventions Against Reward Hacking

This project is an extension of work done for Neel Nanda’s MATS 9.0 Training Phase. Neel Nanda and Josh Engels advised the project. Initial work on this project was done with David Vella Zarb. Thank you to Arya Jakkli, Paul Bogdan, and Monte MacDiarmid for providing feedback on the post...

Dec 29, 202560

Can we interpret latent reasoning using current mechanistic interpretability tools?

Authors: Bartosz Cywinski*, Bart Bussmann*, Arthur Conmy**, Joshua Engels**, Neel Nanda**, Senthooran Rajamanoharan** * primary contributors ** advice and mentorship TL;DR We study a simple latent reasoning LLM on math tasks using standard mechanistic interpretability techniques to see whether the latent reasoning process (i.e., vector-based chain of thought) is interpretable....

Dec 22, 202544

How Can Interpretability Researchers Help AGI Go Well?

Executive Summary * Over the past year, the Google DeepMind mechanistic interpretability team has pivoted to a pragmatic approach to interpretability, as detailed in our accompanying post [1] , and are excited for more in the field to embrace pragmatism! In brief, we think that: * It is crucial to...

Dec 1, 202567

A Pragmatic Vision for Interpretability

Executive Summary * The Google DeepMind mechanistic interpretability team has made a strategic pivot over the past year, from ambitious reverse-engineering to a focus on pragmatic interpretability: * Trying to directly solve problems on the critical path to AGI going well [[1]] * Carefully choosing problems according to our comparative...

Dec 1, 2025136

Current LLMs seem to rarely detect CoT tampering

Authors: Bartosz Cywinski*, Bart Bussmann*, Arthur Conmy**, Neel Nanda**, Senthooran Rajamanoharan**, Joshua Engels** * equal primary contributor, order determined via coin flip ** equal advice and mentorship, order determined via coin flip > “Tampering alert: The thought "I need to provide accurate, helpful, and ethical medical advice" is not my...

Nov 19, 202556