INDEX

Explanations

be the first

6 followed by (

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

staben

0.42

gna

0.40

 protest

0.39

 ofici

0.39

 dieses

0.38

夫

0.37

 órdenes

0.36

 Ngoài

0.36

 প্রদর্শ

0.36

ONI

0.36

POSITIVE LOGITS

be

0.63

 thankful

0.59

 pleased

0.57

Be

0.57

Be

0.54

 thrilled

0.53

 happy

0.52

 begeistert

0.51

 fascinated

0.50

 confident

0.48

Activations Density 0.018%