Truthful LMs as a warm-up for aligned AGI
📖 Description
*This post is heavily informed by prior work, most notably that of Owain Evans, Owen Cotton-Barratt and others (*[*Truthful AI*](https://www.alignmentforum.org/posts/aBixCPqSnTsPsTJBQ/truthful-ai-developing-and-governing-ai-that-does-not-lie)*), Beth Barnes (*[*Risks from AI persuasion*](https://www.alignmentforum.org/posts/5cWtwATHL6KyzChck/risks-from-ai-persuasion)*), Paul Christiano (unpublished) and Dario Amodei (unpublished), but was written by me and is not necessarily endorsed by those people. I am also very grateful to Paul Christiano, Leo Gao, Beth Barnes, William Saunders, Owain Evans, Owen Cotton-Barratt, Holly Mandel and Daniel Ziegler for invaluable feedback.*
📊 Game Impacts
| Variable | Change | Condition |
|---|---|---|
| Research | +5 | Always |
| Vibey Doom | +5 | Always |
| Ethics Risk | -5 | Always |
💭 Reactions
"This is a significant contribution to alignment research"
"Discussed in AI safety community"
🔗 Sources
🤝 Found an Issue?
This event data is sourced from the pdoom-data repository. If you notice errors or want to suggest improvements:
GitHub Issue (Preferred) 📧 Email (No GitHub)