INDEX

Explanations

greetings and friendly salutations

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

__*/

-0.51

########.

-0.48

zzlement

-0.48

wantErr

-0.48

期刊论文

-0.47

DeleteBehavior

-0.47

 entière

-0.47

 fallait

-0.47

 devront

-0.46

[*]

-0.46

POSITIVE LOGITS

 hello

0.83

 greetings

0.77

 Hello

0.76

 greeting

0.76

Greeting

0.75

Hello

0.74

Bonjour

0.72

Greetings

0.72

Hallo

0.71

 Greeting

0.71

Activations Density 0.070%