INDEX

Explanations

numbers, citations, and programming code, and it often picks up nearby words.

the number 7

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

Negative Logits

исленность

-0.84

 resourceCulture

-0.79

FormTagHelper

-0.78

NUMX

-0.73

WireFormatLite

-0.73

)++;

-0.72

 houſe

-0.70

 pleaſure

-0.69

TypedDataSet

-0.68

Tembelea

-0.68

POSITIVE LOGITS

 உள்ள

0.46

alt

0.45

poz

0.45

Poz

0.45

hnen

0.43

gli

0.42

meme

0.42

 phấn

0.40

rep

0.40

替代

0.38

Activations Density 10.940%