INDEX

Explanations

#define

This neuron fires on programming preprocessor directives—especially “#include” and “#define” lines.

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

of

-0.92

 From

-0.85

 Accesat

-0.85

 individuale

-0.84

 prestigio

-0.82

Ｊ

-0.81

④

-0.81

 poignant

-0.80

 Vordergrund

-0.78

 first

-0.77

POSITIVE LOGITS

 vêtements

1.08

esie

1.06

 creativos

1.02

 olvidar

0.99

 trastornos

0.96

 decoraciones

0.96



0.94

 protectora

0.94

 actuales

0.94

 frukt

0.93

Activations Density 0.049%