Notes on the Alignment Forum's Value Learning sequence curated by Rohin Shah. ambitious value learning : the idea of learning 'the human…
[ grokking ] / [ phase change hypothesis ] emergence of near-discrete features in large transformers symmetries / non-[ identifiable…
I think of "variance" as the error in a statistical estimate that comes from not having enough data (assuming an [ identifiable ] model…
Philosophical views on consciousness: Buddhist and meditative traditions focus on [ awareness ]. They claim that consciousness has nothing…
The Bayesian approach to statistics is to 'just use probability theory'. You write down a joint probability distribution over observed and…