INDEX

Explanations

wrong, unethical, disrespectful, problematic

New Auto-Interp

Configuration

Prompts (Dashboard)

238,145 prompts, 512 tokens each

Dataset (Dashboard)

lmsys + oasst1

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 pressured

0.71

 influenz

0.71

 🙂

0.70

ุงเทพ

0.70

 Advantage

0.67

 risky

0.66

 kesulitan

0.66

 advantage

0.66

 منفی

0.65

보다는

0.63

POSITIVE LOGITS

 abhor

1.33

 heinous

1.33

 egregious

1.30

 abomin

1.29

 violation

1.28

 affront

1.24

 atrocious

1.21

 outrage

1.21

 disgraceful

1.19

 disgrace

1.18

Activations Density 0.380%