INDEX

Explanations

alerts and code blocks

The neuron strongly activates on HTML element names and CSS class tokens—especially UI/Bootstrap component keywords (e.g. “alert,” “badge,” “progress,” “jumbo,” etc.).

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

The

-1.34

who

-1.21

-1.16

),

-1.16

 These

-1.15

に入れ

-1.14

int

-1.13

 This

-1.11

…)

-1.10

 fünf

-1.09

POSITIVE LOGITS

Fonto

1.29

延伸閱讀

1.27

بسم

1.25

 variantes

1.21

 organik

1.20

fok

1.20

بله

1.17

 vallen

1.16

Cinta

1.16

 legger

1.15

Activations Density 0.002%