INDEX

Explanations

classification or category

This neuron strongly activates on programming-related identifiers and API names—especially .NET framework class, namespace, and other code tokens.

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

 CentOS

-0.81

patra

-0.74

 Blackwood

-0.72

ρεί

-0.72

 förstå

-0.71

cayaan

-0.70

 Prosper

-0.70

wiki

-0.70

╤

-0.70

 festgestellt

-0.69

POSITIVE LOGITS

 design

0.89

 category

0.87

category

0.84

 kategori

0.81

 катего

0.80

 Weld

0.79

uchsia

0.76

 editor

0.72

 Category

0.72

 weld

0.71

Activations Density 0.021%