The Redundant Information Hypothesis

Created time
Mar 17, 2023 04:41 PM
Main Box
Tags
Natural Abstractions
Interpretability
AI Alignment
Public

The Redundant Information Hypothesis

The RIH basically says that natural abstractions can be well described by mathematical functions of redudant or conserved information.

Source: Chan, L., Lang, L. and Jenner, E. (no date) ‘Natural Abstractions: Key claims, Theorems, and Critiques’. Available at: https://www.alignmentforum.org/posts/gvzW46Z3BsaZsLc25/natural-abstractions-key-claims-theorems-and-critiques-1 (Accessed: 19 March 2023).