INDEX

Explanations

errors and code snippets

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

kär

-1.33

 Usar

-1.32

 Keramik

-1.30

 delicada

-1.30

が必要です

-1.27

 cheia

-1.27

 perfeita

-1.27

ほら

-1.24

 payé

-1.24

のもの

-1.23

POSITIVE LOGITS

 after

1.53

by

1.33

 since

1.30

 while

1.29

for

1.29

 which

1.27

 during

1.26

了许多

1.21

了不少

1.17

 just

1.16

Activations Density 0.008%