The Universality Hypothesis

Created time
Mar 17, 2023 04:39 PM
Tags
Natural Abstractions
AI Alignment
Interpretability
Public

The Universality Hypothesis

The Natural Abstractions Hypothesis usually gets separated into two seperate claims:
The Universality Hypothesis and
The Redundant Information Hypothesis
.
The Universality Hypothesis
basically claims that natural abstractions exist, because many intelligent agents learn similar abstractions.

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).