INDEX

Explanations

code identifiers and types

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

 bedre

-1.42

っていたので

-1.35

 nastro

-1.32



-1.32

 beſt

-1.30

 vijf

-1.29

 firmy

-1.27

jeg

-1.27

 zespo

-1.26

wendig

-1.23

POSITIVE LOGITS

If

1.60

Following

1.49

There

1.47

1.41

 verstehen

1.41

!!!!!

1.40

!!!!

1.38

At

1.37

For

1.34

 ostensibly

1.33

Activations Density 0.015%