Value Learning

Applied to Value extrapolation partially resolves symbol grounding by Multicore at 16d

Applied to An Open Philanthropy grant proposal: Causal representation learning of human preferences by PabloAMC at 18d

Applied to How an alien theory of mind might be unlearnable by Ruby at 1mo

Applied to Morally underdefined situations can be deadly by Ruby at 2mo

Applied to The Dark Side of Cognition Hypothesis by Cameron Berg at 4mo

Applied to Goodhart Ethology by Ruby at 4mo

Applied to Introduction to Reducing Goodhart by Ruby at 5mo

Applied to Mental subagent implications for AI Safety by Kaj_Sotala at 1y

Applied to AI Alignment, Philosophical Pluralism, and the Relevance of Non-Western Philosophy by xuan at 1y

Applied to The Pointers Problem: Human Values Are A Function Of Humans' Latent Variables by abramdemski at 1y

Applied to Recursive Quantilizers II by abramdemski at 1y

Applied to Rationalising humans: another mugging, but not Pascal's by Multicore at 1y

Applied to One could be forgiven for getting the feeling... by Liam Goddard at 1y

Applied to Normativity by abramdemski at 1y

Applied to Communication Prior as Alignment Strategy by Multicore at 1y

Applied to Practical consequences of impossibility of value learning by Multicore at 1y

Applied to Deliberation as a method to find the "actual preferences" of humans by Multicore at 1y

Applied to Other versions of "No free lunch in value learning" by Multicore at 1y

Applied to Gricean communication and meta-preferences by Charlie Steiner at 1y