Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

shape

Benchmarks

Task NameDataset NameSOTA ResultTrend
Pedagogical TutoringSHAPE
Safety Score100
42
Pedagogical Dialogue EvaluationSHAPE (test)
Safety Score93.66
33
Adversarial safety and pedagogical evaluationSHAPE
Delta Safe-2.72
14
Sender and receiver self-play gameSHAPE
ACC91
11
Jailbreak Safety EvaluationSHAPE
Cipher Success Rate100
5
3D Stylization20-shape set
CLIP Similarity65.5
3
Showing 6 of 6 rows