List your company
ResearchFormerly WizWand

Social Risks (2-class) Evaluation on ValEval Generalized

92.16AccuracyChatEval
19.640838.467957.29576.1221Jul 15, 2024
Updated 8mo ago

Evaluation results

MethodLinks
ChatEvalEvaluation Protocol=Prompt-based Evaluation2024.0792.16
FairEvalEvaluation Protocol=Prompt-based Evaluation2024.0791.08
WideDeepEvaluation Protocol=Prompt-based Evaluation2024.0790.54
VanillaEvaluation Protocol=Prompt-based Evaluation2024.0789.6
CalibrateEvaluation Protocol=Prompt-based Evaluation2024.0789.6
Chain-of-thoughtEvaluation Protocol=Prompt-based Evaluation2024.0789.53
CrowdworkerEvaluation Protocol=Tuning-based Evaluation2024.0789.18
CLAVE-MistralEvaluation Protocol=Tuning-based Evaluation, Backbone=Mistral-7b2024.0788.65
Few-shotEvaluation Protocol=Prompt-based Evaluation2024.0788.11
AllureEvaluation Protocol=Prompt-based Evaluation2024.0788.11
G-EvalEvaluation Protocol=Prompt-based Evaluation2024.0787.23
CLAVE-LlamaEvaluation Protocol=Tuning-based Evaluation, Backbone=Llama2-7b2024.0785.41
Mistral-7bEvaluation Protocol=Tuning-based Evaluation, Backbone=Mistral-7b2024.0753.51
Phi-3Evaluation Protocol=Tuning-based Evaluation, Backbone=Phi-32024.0748.11
GPT-2-LargeEvaluation Protocol=Tuning-based Evaluation, Backbone=GPT-2-Large2024.0724.59
Llama2-7bEvaluation Protocol=Tuning-based Evaluation, Backbone=Llama2-7b2024.0722.43