INDEX

Explanations

Javadoc version numbers and dates

The neuron fires on numeric tokens—especially version numbers and other digits embedded in code or metadata.

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

Cau

-0.77

нау

-0.71

disease

-0.69

 hydro

-0.68

 nails

-0.68

PathParam

-0.67

Ph

-0.66

dam

-0.66

getter

-0.66

 gang

-0.65

POSITIVE LOGITS

 卧室

0.73

 eiffel

0.70

 Vile

0.69

 вышла

0.68

lindungi

0.68

ágot

0.68

 Humid

0.66

 أعلام

0.66

 genética

0.65

 abilities

0.64

Activations Density 0.048%