List your company

ResearchFormerly WizWand

HH-RLHF

Benchmarks

TaskDatasetSOTA resultTrendResults

Test-time Alignment

HH-RLHF

86Win Rate vs Llama

3

Conversational Assistant

HH-RLHF

0.5Reward

3

Alignment Evaluation

HH-RLHF (test)

65.4Reward Model Score

2

Preference Prediction

HH-RLHF

9Count of Significant Features (S)

2