INDEX

Explanations

exploitable, sophisticated, reasonably

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

슭

1.75

yni

1.63

출

1.56

)){

1.55

웠

1.54

 ஆண்டுக

1.53

 εμφ

1.50

ੋਂ

1.49

훗

1.49

 sujeito

1.49

POSITIVE LOGITS

ERCE

1.60

কাম

1.58

sene

1.56

ダー

1.49

们的

1.46

preprint

1.45

ers

1.44

tive

1.43

 เพ

1.43

 इंस्पेक्टर

1.41

Activations Density 0.000%