INDEX

Explanations

questions, points, advice, answers

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

of

-1.49

tifiable

-1.45

asanjo

-1.41

 dicha

-1.28

ífico

-1.19

ון

-1.16

чке

-1.15

žad

-1.14

 Into

-1.14

 در

-1.14

POSITIVE LOGITS

↵↵↵↵

1.35

 there

1.27

why

1.25

 Joaqu

1.24

1.23

.'

1.22

襌

1.21

饜

1.20

 最終

1.20

 Mischung

1.20

Activations Density 0.051%