INDEX

Explanations

uppercase letters

The neuron fires on individual uppercase letters (A–Z) in regex character classes.

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

);

-2.02

by

-1.73

一旁的

-1.70

 getting

-1.70

各大

-1.66

 delivering

-1.60

 using

-1.59

if

-1.59

 subjected

-1.56

any

-1.55

POSITIVE LOGITS

“

2.06

籀

1.97

1.90

および

1.77

цией

1.75

 insanely

1.72

镦

1.65

 esche

1.65

𐀀

1.63

 abhor

1.61

Activations Density 0.002%