Playground 2 — Label Preference Data

You are the human annotator. Click the response you prefer — watch the reward model's training loss decrease as you add labels.

Prompt
Labels collected
0
RM training loss (simulated)