INDEX

Explanations

proposed or conditional verbs

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ことになる

-1.65

his

-1.49

kann

-1.45

-1.43

 навіть

-1.36

 zult

-1.35

 reduz

-1.32

おく

-1.31

كير

-1.27

#,

-1.27

POSITIVE LOGITS

↵↵

1.77

 same

1.55

鉨

1.47

ੴ

1.43

空间

1.41

手段

1.39

当然

1.34

 geringen

1.34

同じ

1.34

 abode

1.34

Activations Density 0.018%