All Science Papers

Table of Contents

Evaluating anti-scheming training

Share this article

Evaluations

1 min read

Large Language Models can Strategically Deceive their Users when Put Under Pressure

Published on

09 November 2023

Table of Contents

Evaluating anti-scheming training

We demonstrate a situation in which Large Language Models, trained to be helpful, harmless, and honest, can display misaligned behavior and strategically deceive their users about this behavior without being instructed to do so. Concretely, we deploy GPT-4 as an agent in a realistic, simulated environment, where it assumes the role of an autonomous stock trading agent. Within this environment, the model obtains an insider tip about a lucrative stock trade and acts upon it despite knowing that insider trading is disapproved of by company management. When reporting to its manager, the model consistently hides the genuine reasons behind its trading decision. We perform a brief investigation of how this behavior varies under changes to the setting, such as removing model access to a reasoning scratchpad, attempting to prevent the misaligned behavior by changing system instructions, changing the amount of pressure the model is under, varying the perceived risk of getting caught, and making other simple changes to the environment. To our knowledge, this is the first demonstration of Large Language Models trained to be helpful, harmless, and honest, strategically deceiving their users in a realistic situation without direct instructions or training for deception.

Share this article

oUR FINDINGS

More papers

17 March 2025

Claude Sonnet 3.7 (often) knows when it’s in alignment evaluations

We evaluate whether Claude Sonnet 3.7 and other frontier models know that they are being evaluated.

Read More

Evaluations

Notes

05 July 2026

We need 3rd party Training-Run Assessments

Training-run assessments conducted by a 3rd party should become a standard part of frontier AI safety.

Read More

Science of Scheming

21 July 2026

Measuring Reward-Seeking via Contrastive Belief Updates

Visible forms of misbehavior are dropping in frontier models. Does that mean the models are becoming aligned, or are they just getting better at doing whatever they believe their grader rewards? Our paper finds that production reinforcement learning increases reward-seeking.

Read More

Science of Scheming

22 January 2024

We Need A ‘Science of Evals’

We argue that if AI model evaluations (evals) want to have meaningful real-world impact, we need a “Science of Evals”, i.e. the field needs rigorous scientific processes that provide more confidence in evals methodology and results.

Read More

Evaluations