INDEX

Explanations

generated javadoc comments

This neuron detects references to base-64 encoding (e.g. “base64” imports or function calls).

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

ppm

-0.79

 nemoc

-0.78

 TextStyle

-0.77

 Diabetic

-0.75

لمين

-0.75

 Mineral

-0.75

颜

-0.74

diagnos

-0.73

 Bitmap

-0.72

 insulin

-0.71

POSITIVE LOGITS

ongeza

0.87

白

0.74

 urls

0.74

urz

0.73

urls

0.71

贝

0.71

 Panda

0.70

Sorry

0.69

 Dewey

0.69

 envelop

0.69

Activations Density 0.008%