INDEX

Explanations

Wolfram Mathematica Docket

The neuron selectively responds to numeric tokens (numbers and digit sequences) in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 email

-1.02

 emailed

-0.84

TextLabel

-0.78

傾向

-0.78

ixen

-0.76

ByEmail

-0.75

 fprintf

-0.74

 emailing

-0.72

sendEmail

-0.72

 gmail

-0.71

POSITIVE LOGITS

okolade

0.80

名字

0.75

PackageManager

0.75

 Statements

0.75

ﾋ

0.73

tanque

0.73

 LUCK

0.72

LIO

0.72

 PICTURE

0.72

phenol

0.71

Activations Density 0.006%