INDEX

Explanations

programming constructs

The neuron fires on code identifiers—especially function and type names—that begin with a capital letter.

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

嚀

-1.13

齜

-1.13

 リボン

-1.04

время

-1.02

/*!

-1.02

ঈ

-1.00

 özellikleri

-0.99

Nike

-0.98

哙

-0.98

 vijf

-0.97

POSITIVE LOGITS

by

1.67

of

1.49

 these

1.39

in

1.38

 from

1.28

at

1.20

 within

1.16

1.05

 during

1.05

 before

1.01

Activations Density 0.007%