INDEX

Explanations

HTML tag "<"

This neuron has a diverse set of activations, and appears to be firing on section headers, list items, article titles and other document structure elements

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

 doInBackground

-0.82

InjectAttribute

-0.81

 kasarigan

-0.79

 Wicidata

-0.77

principalTable

-0.71

ftagPool

-0.69

 Lightboxes

-0.69

JspWriter

-0.67

EditorBrowsable

-0.65

 AssemblyCulture

-0.64

POSITIVE LOGITS

sie

0.45

نمای

0.43

Zim

0.43

電

0.41

Third

0.41

<table>

0.41

 Third

0.41

LineEdit

0.40

 Sixt

0.40

 merry

0.40

Activations Density 2.021%