INDEX

Explanations

markdown formatting characters

This neuron activates on markup/formatting tokens (e.g. Markdown asterisks, backticks, and HTML tags) used to denote emphasis or styling.

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

 baumwolle

-1.43

휀

-1.42

鰨

-1.40

ishw

-1.39

 gorra

-1.38

 fonctionnalité

-1.37

Laguna

-1.36



-1.35

Studios

-1.35

 ファイル

-1.35

POSITIVE LOGITS

 from

1.85

↵↵

1.59

These

1.57

This

1.52

mathrm

1.50

Our

1.46

1.43

Among

1.43

All

1.42

Along

1.41

Activations Density 0.003%