INDEX

Explanations

in

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

<bos>

-0.52

 winter

-0.46

 onely

-0.46

ToTensor

-0.45

 drugs

-0.45

wissenschaften

-0.44

 disappointing

-0.44

 kuto

-0.44

 Spitz

-0.43

 violent

-0.42

POSITIVE LOGITS

 <<<<<<<<<<<<<<

0.75

[]:

0.67

 שוליים

0.66

الإنجليزية

0.60

Sucesor

0.59

liothèque

0.58

/\.

0.56

ContentAsync

0.55

Jegyzetek

0.54

Lähteet

0.54

Activations Density 0.599%

No Known Activations