evaluation-task · Source-linked discovery

The Art of Saying No: Contextual Noncompliance in Language Models

Dataset with 1001 samples to test noncompliance capabilities of language models. Contrast set of 379 samples.

Open interactive record →
OriginFaeze Brahman, Sachin Kumar, Vidhisha Balachandran et al.TopicsSafeguardsStatusimported

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.

Original sources