Longterm Wiki

Eliciting Latent Knowledge (ELK)

Alignment Research CenterResearch Areaseliciting-latent-knowledge-elk

Record Metadata

Record Keyeliciting-latent-knowledge-elk
EntityAlignment Research Center
CollectionResearch Areas(1 record total)
SchemaMajor research initiatives and focus areas.
YAML Filepackages/kb/data/things/QsXVXtQ0zE.yaml

Fields

NameEliciting Latent Knowledge (ELK)
DescriptionResearch on extracting truthful knowledge from AI models regardless of learned deceptive behaviors
StartedDec 2021
Key Publicationdocs.google.com
NotesSeminal ELK report posed as a prize problem; attracted significant community engagement
Record: eliciting-latent-knowledge-elk | Longterm Wiki