ELK: Eliciting Latent Knowledge
GPTKB entity
Statements (16)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:academic_journal
gptkb:AI_safety_proposal |
| gptkbp:abbreviation |
gptkb:ELK
|
| gptkbp:author |
gptkb:ARC_(Alignment_Research_Center)
gptkb:Paul_Christiano |
| gptkbp:focusesOn |
AI alignment
eliciting knowledge from AI models latent knowledge |
| gptkbp:problemAddressed |
AI models may know things they cannot or will not say
|
| gptkbp:proposedBy |
methods to extract knowledge from AI systems
|
| gptkbp:publicationYear |
2021
|
| gptkbp:publishedBy |
gptkb:Alignment_Research_Center
|
| gptkbp:url |
https://www.alignmentforum.org/posts/6z3A4E9Zr9L5oG9zA/eliciting-latent-knowledge
|
| gptkbp:bfsParent |
gptkb:Paul_Christiano
|
| gptkbp:bfsLayer |
6
|
| https://www.w3.org/2000/01/rdf-schema#label |
ELK: Eliciting Latent Knowledge
|