INDEX

Explanations

name and id

The neuron is primarily detecting colon characters (":"), especially when they introduce definitions, list headers, or field labels.

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

 those

-1.03

nergie

-0.93

and

-0.92

ناسب

-0.90

DirPath

-0.79

 sejauh

-0.79

any

-0.79

corpo

-0.77

 resolution

-0.77

صبحت

-0.77

POSITIVE LOGITS

 name

1.20

廢

1.03

ょう

0.91

()!=

0.88

 carré

0.86

 découvrez

0.83

мә

0.82

%;

0.81

?,?,

0.81

 üre

0.80

Activations Density 0.070%