/explore
|
You are here 
|
www.lesswrong.com | ||
| | | | |
www.danieldjohnson.com
I argue that hallucinations are a natural consequence of the language modeling objective, which focuses on simulating confident behavior even when that behavior is hard to predict, rather than predictable behaviors that take uncertainty into account. I also discuss five strategies for avoiding this mismatch. |
|
| | | | | ||
| | | | |
enbao.me
toxic data, alignment faking, dynamic chunking, entropy-based exploration, general reward models, magistral |
|
| | | | | ||
| | | | |
www.alignmentforum.org
This is the third of seven posts in the Conditioning Predictive Models Sequence based on the paper "Conditioning Predictive Models: Risks and Strateg... |
|
| | | | | ||
| | | | |
www.lesswrong.com
(Independent) alignment researchers need to be exceptionally good at philosophy of science In order to be effective, independent AI alignment/safety/... |
|
| | | |||
