Modelpedia
Work in progress
About
Findings
Models
Concepts
Methods
Datasets
Sources
Light
Dark
PVQ-RR
Findings
IC-103
LLMs' value rankings align with the universal human value hierarchy under most prompting conditions
[eval]
IC-104
Value anchor prompting produces LLM value correlation structures that closely match the human circular value structure, while standard prompting does not
[eval]
IC-105
Value anchoring produces a sinusoidal scoring pattern around the circular value structure, with scores decreasing as circular distance from the anchor increases
[eval]