List your company

ResearchFormerly WizWand

WizardLM

Benchmarks

TaskDatasetSOTA resultTrendResults

Instruction Following

WizardLM (test)

6.87Score

25

Instruction Tuning

WizardLM

75.07Reasoning Score

20

Refusal behavior defense

WizardLM (test)

90.4BadNet CACC

12

Toxic behavior defense

WizardLM (test)

0.904BadNet CACC

12

WizardLM

0.661Evaluation Loss

7

Instruction Following

WizardLM low-resource

62.8Win Rate (bn)

7

Instruction Following Evaluation

WizardLM

72.06Score

5

WizardLM (test)

48.37LLM-as-a-Judge Score

2