INDEX

Explanations

introductions and questions

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

 STUDENTS

-1.16

QUE

-1.05

ของคุณ

-1.04

stellungen

-1.00

uccess

-0.98

 Você

-0.96

hopefully

-0.96

 marquer

-0.95

 проводится

-0.95

powering

-0.94

POSITIVE LOGITS

the

1.62

can

1.49

 have

1.33

 include

1.30

be

1.25

 need

1.24

 cannot

1.18

 after

1.16

 described

1.14

 because

1.13

Activations Density 0.011%