INDEX

Explanations

programming and localization

This neuron fires on standalone labels or headings—single-word commands, menu items, or section titles.

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

Even

-1.45

 padx

-1.37

Whether

-1.36

可以

-1.27

我要

-1.25

 -----

-1.25

Pendant

-1.25

、

-1.24

où

-1.23

 даже

-1.23

POSITIVE LOGITS

of

2.25

 among

1.75

at

1.71

to

1.70

 with

1.68

 around

1.48

umplimiento

1.38

 into

1.26

 asmen

1.25

 imageName

1.24

Activations Density 0.074%