INDEX

Explanations

specific categories and their characteristics

The neuron responds to domain-specific content nouns—especially plural or mass-noun labels of key topics, objects, or research elements.

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

 素描

-1.70

俢

-1.57

dzić

-1.55

对不起

-1.54

苼

-1.50

都是

-1.49

儈

-1.46

 arquitetura

-1.45

现在的

-1.45

acakt

-1.45

POSITIVE LOGITS

 that

2.81

 those

2.05

2.02

 these

1.99



1.82

 tiga

1.80

 this

1.76

 such

1.74

:.

1.73

 veya

1.72

Activations Density 0.363%