INDEX

Explanations

The neuron strongly fires on closing-parenthesis tokens ( “)” ) typically found in code or technical markup.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ळख

0.77

रातील

0.75

खाव

0.71

णाची

0.69

विण्यासाठी

0.64

aménagement

0.63

 प्रकारचे

0.63

 चळ

0.63

decorative

0.63

 कामगिरी

0.63

POSITIVE LOGITS

1.69

{//

1.27

){

1.13

{}

1.08

(){

1.03

){

0.97

 throws

0.96

0.93

{})

0.90

{.

0.89

Activations Density 0.001%