LESSWRONG
LW

Hypothesis Subspace

Sep 12, 2022 by Paul Bricman

A living collection of alignment proposals I'm exploring at Refine, a program hosted by Conjecture.

20Oversight Leagues: The Training Game as a Feature

3y

6

6Ideological Inference Engines: Making Deontology Differentiable*

3y

0

30Representational Tethers: Tying AI Latents To Human Ones

3y

0

15Interlude: But Who Optimizes The Optimizer?

3y

0

25(Structural) Stability of Coupled Optimizers

3y

0

13Cataloguing Priors in Theory and Practice

3y

8