INDEX

Explanations

HTML `select` tags

This neuron fires on all‐caps technical acronyms or code‐related tokens (e.g. HTML, CSS, SQL, etc.).

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

鄀

0.66

('/:

0.63

 pickle

0.59

 kutoka

0.57

 lookup

0.56

 glacier

0.55

Ack

0.54

 मुर

0.54

 unripe

0.54

 coalgebra

0.54

POSITIVE LOGITS

 HTML

1.65

HTML

1.63

CSS

1.52

 XHTML

1.48

 html

1.47

Html

1.42

CSS

1.41

html

1.35

 divs

1.33

 Html

1.31

Activations Density 0.183%