INDEX

Explanations

on the other hand

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

and

-2.02

でしたが

-1.55

éndose

-1.41

總結

-1.37

してます

-1.37

 فره

-1.36

 rezultat

-1.32

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

-1.27

 feme

-1.27

archiwizowane

-1.26

POSITIVE LOGITS

1.85

it

1.60

are

1.57

 conversely

1.51

1.39

can

1.38

*}{

1.38

1.35

 ความ

1.28

 universitaria

1.27

Activations Density 0.008%