AI Safety Fundamentals: Alignment

83 Avsnitt

1. Is Power-Seeking AI an Existential Risk?

   Publicerades: 2023-05-13
2. Where I Agree and Disagree with Eliezer

   Publicerades: 2023-05-13
3. Supervising Strong Learners by Amplifying Weak Experts

   Publicerades: 2023-05-13
4. Measuring Progress on Scalable Oversight for Large Language Models

   Publicerades: 2023-05-13
5. Least-To-Most Prompting Enables Complex Reasoning in Large Language Models

   Publicerades: 2023-05-13
6. Summarizing Books With Human Feedback

   Publicerades: 2023-05-13
7. Takeaways From Our Robust Injury Classifier Project [Redwood Research]

   Publicerades: 2023-05-13
8. AI Safety via Debatered Teaming Language Models With Language Models

   Publicerades: 2023-05-13
9. High-Stakes Alignment via Adversarial Training [Redwood Research Report]

   Publicerades: 2023-05-13
10. AI Safety via Debate

    Publicerades: 2023-05-13
11. Robust Feature-Level Adversaries Are Interpretability Tools

    Publicerades: 2023-05-13
12. Introduction to Logical Decision Theory for Computer Scientists

    Publicerades: 2023-05-13
13. Debate Update: Obfuscated Arguments Problem

    Publicerades: 2023-05-13
14. Discovering Latent Knowledge in Language Models Without Supervision

    Publicerades: 2023-05-13
15. Feature Visualization

    Publicerades: 2023-05-13
16. Toy Models of Superposition

    Publicerades: 2023-05-13
17. Understanding Intermediate Layers Using Linear Classifier Probes

    Publicerades: 2023-05-13
18. Acquisition of Chess Knowledge in Alphazero

    Publicerades: 2023-05-13
19. Careers in Alignment

    Publicerades: 2023-05-13
20. Embedded Agents

    Publicerades: 2023-05-13