INDEX

Explanations

using

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 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

 amargo

-0.50

__":

-0.50

 angusti

-0.49

 facteur

-0.48

Prophet

-0.48

taxia

-0.47

 Pupils

-0.47

readystatechange

-0.45

 Inflate

-0.45

 tungkol

-0.45

POSITIVE LOGITS

حياته

0.65

AndEndTag

0.64

}{*}{

0.62

ParallelGroup

0.59

XmlAccessorType

0.58

<bos>

0.56

حياتها

0.56

fxml

0.55

 manip

0.55

 think

0.55

Activations Density 0.007%