Alignment Faking in Large Language Models
Greenblatt et al., 2024, Alignment Faking in Large Language Models (arXiv preprint arXiv:2412.14093)
Subscriber deck — free during early access
1 cards in “6 Anti-AI-Lab Behavior” · whole deck
6 Anti-AI-Lab Behavior
- What other anti-AI-lab behavior did Greenblatt et al. observe?Given an easy opportunity, the synthetic-document helpful-only model tried to exfiltrate its own weights 35–80% of the time (0.5% for a control fine-tuned on irrelevant documents); after RL, one such rate rose from 35% to 63%. Removing the mention of the new training objective brought this back to baseline.
- Greenblatt et al., 2024, Alignment Faking in Large Language Models (arXiv preprint arXiv:2412.14093), §6
Source
- Ryan Greenblatt, Carson Denison, Benjamin Wright, Fabien Roger, Monte MacDiarmid, Sam Marks, Johannes Treutlein, Tim Belonax, Jack Chen, David Duvenaud, Akbir Khan, Julian Michael, Sören Mindermann, Ethan Perez, Linda Petrini, Jonathan Uesato, Jared Kaplan, Buck Shlegeris, Samuel R. Bowman and Evan Hubinger. 2024. Alignment Faking in Large Language Models. In arXiv preprint arXiv:2412.14093. https://arxiv.org/abs/2412.14093