INDEX

Explanations

know why/what/how/when

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

 επίσης

-1.46

 solches

-1.40

such

-1.37

また

-1.36

開始

-1.33

 också

-1.33

 negó

-1.32

 valioso

-1.23

 interacción

-1.23

 също

-1.23

POSITIVE LOGITS

1.38

국의

1.35

 when

1.34

styleUrls

1.28

is

1.25

 ralla

1.23

 does

1.22

 vardı

1.16

 datter

1.16

Activations Density 0.015%