AXRP - the AI X-risk Research Podcast

39 - Evan Hubinger on Model Organisms of Misalignment


Listen Later

The 'model organisms of misalignment' line of research creates AI models that exhibit various types of misalignment, and studies them to try to understand how the misalignment occurs and whether it can be somehow removed. In this episode, Evan Hubinger talks about two papers he's worked on at Anthropic under this agenda: "Sleeper Agents" and "Sycophancy to Subterfuge".

Patreon: https://www.patreon.com/axrpodcast

Ko-fi: https://ko-fi.com/axrpodcast

The transcript: https://axrp.net/episode/2024/12/01/episode-39-evan-hubinger-model-organisms-misalignment.html

 

Topics we discuss, and timestamps:

0:00:36 - Model organisms and stress-testing

0:07:38 - Sleeper Agents

0:22:32 - Do 'sleeper agents' properly model deceptive alignment?

0:38:32 - Surprising results in "Sleeper Agents"

0:57:25 - Sycophancy to Subterfuge

1:09:21 - How models generalize from sycophancy to subterfuge

1:16:37 - Is the reward editing task valid?

1:21:46 - Training away sycophancy and subterfuge

1:29:22 - Model organisms, AI control, and evaluations

1:33:45 - Other model organisms research

1:35:27 - Alignment stress-testing at Anthropic

1:43:32 - Following Evan's work

 

Main papers:

Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training: https://arxiv.org/abs/2401.05566

Sycophancy to Subterfuge: Investigating Reward-Tampering in Large Language Models: https://arxiv.org/abs/2406.10162

 

Anthropic links:

Anthropic's newsroom: https://www.anthropic.com/news

Careers at Anthropic: https://www.anthropic.com/careers

 

Other links:

Model Organisms of Misalignment: The Case for a New Pillar of Alignment Research: https://www.alignmentforum.org/posts/ChDH335ckdvpxXaXX/model-organisms-of-misalignment-the-case-for-a-new-pillar-of-1

Simple probes can catch sleeper agents: https://www.anthropic.com/research/probes-catch-sleeper-agents

Studying Large Language Model Generalization with Influence Functions: https://arxiv.org/abs/2308.03296

Stress-Testing Capability Elicitation With Password-Locked Models [aka model organisms of sandbagging]: https://arxiv.org/abs/2405.19550

 

Episode art by Hamish Doodles: hamishdoodles.com

...more
View all episodesView all episodes
Download on the App Store

AXRP - the AI X-risk Research PodcastBy Daniel Filan

  • 4.4
  • 4.4
  • 4.4
  • 4.4
  • 4.4

4.4

8 ratings


More shows like AXRP - the AI X-risk Research Podcast

View all
Making Sense with Sam Harris by Sam Harris

Making Sense with Sam Harris

26,371 Listeners

Conversations with Tyler by Mercatus Center at George Mason University

Conversations with Tyler

2,426 Listeners

a16z Podcast by Andreessen Horowitz

a16z Podcast

1,083 Listeners

Future of Life Institute Podcast by Future of Life Institute

Future of Life Institute Podcast

107 Listeners

The Daily by The New York Times

The Daily

112,430 Listeners

Practical AI by Practical AI LLC

Practical AI

210 Listeners

All-In with Chamath, Jason, Sacks & Friedberg by All-In Podcast, LLC

All-In with Chamath, Jason, Sacks & Friedberg

9,823 Listeners

Machine Learning Street Talk (MLST) by Machine Learning Street Talk (MLST)

Machine Learning Street Talk (MLST)

89 Listeners

Dwarkesh Podcast by Dwarkesh Patel

Dwarkesh Podcast

489 Listeners

Hard Fork by The New York Times

Hard Fork

5,477 Listeners

Clearer Thinking with Spencer Greenberg by Spencer Greenberg

Clearer Thinking with Spencer Greenberg

132 Listeners

The Ezra Klein Show by New York Times Opinion

The Ezra Klein Show

16,106 Listeners

Latent Space: The AI Engineer Podcast by swyx + Alessio

Latent Space: The AI Engineer Podcast

97 Listeners

This Day in AI Podcast by Michael Sharkey, Chris Sharkey

This Day in AI Podcast

209 Listeners

Complex Systems with Patrick McKenzie (patio11) by Patrick McKenzie

Complex Systems with Patrick McKenzie (patio11)

133 Listeners