Investigator models trained with an iterative Frank-Wolfe objective discover diverse, interpretable natural-language prompts that elicit targeted jailbreaks, hallucinations, and aberrant behaviors.
Diagnostic and Statistical Manual of Mental Disorders (DSM-5 )
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
dataset 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
dataset 1polarities
use dataset 1representative citing papers
citing papers explorer
-
Eliciting Language Model Behaviors with Investigator Agents
Investigator models trained with an iterative Frank-Wolfe objective discover diverse, interpretable natural-language prompts that elicit targeted jailbreaks, hallucinations, and aberrant behaviors.