evaluation-task · Source-linked discovery
BOLD: Bias in Open-ended Language Generation Dataset
A dataset to measure fairness in open-ended text generation, covering five domains: profession, gender, race, religious ideologies, and political ideologies.
OriginJwala Dhamala, Tony Sun, Varun Kumar et al.TopicsGeneral capabilityStatusimported
Can support
Not independently assessed by FronteraEval yet.
Cannot support by itself
No inference beyond the upstream source should be made until the protocol is reviewed.