INDEX

Explanations

opening"},"first":1.0,"last":1.0,"text":"The neuron looks for patterns that indicate an opening or introduction to a topic or list. These often start with specific transition words."}

## headings and markdown

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

："

0.48

जापुर

0.44

:");

0.44

LEXPORT

0.43

hygine

0.43

:"

0.43

:"

0.41

:\

0.40

 antib

0.40

：“

0.40

POSITIVE LOGITS

<strong>

0.68

<h3>

0.58

<h5>

0.57

<h4>

0.49

<em>

0.48

[/

0.48

0.46

<b>

0.46

if

0.43

[]){

0.41

Activations Density 0.023%