INDEX

Explanations

understanding user feelings or reasons

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

 directing

1.38

 bitter

1.33

fon

1.33

 treasures

1.27

 mansion

1.25

wartz

1.25

 television

1.24

 tyrann

1.22

 bald

1.21

 Superintend

1.21

POSITIVE LOGITS

።

1.59

ச்சி

1.53

ﻲ

1.49

🄴

1.45

integral

1.38

 اقتص

1.38

тия

1.36

 окно

1.36

 envio

1.33

 pacote

1.32

Activations Density 0.061%