INDEX

Explanations

This neuron fires on markup elements and attributes—i.e. HTML/JSX tag names (like “div”, “main”) and attribute names (like “class”, “id”) in code.

html div or View tags

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

its

-1.45

 Each

-1.40

but

-1.27

 However

-1.24

 Also

-1.20

Its

-1.17

 These

-1.16

 about

-1.15

to

-1.15

 first

-1.14

POSITIVE LOGITS

缡

1.33

🔰

1.21

峮

1.16

 rése

1.15

 geni

1.11

 appel

1.09

 fils

1.08

艄

1.08

 pijama

1.05

 cellule

1.05

Activations Density 0.004%