INDEX

Explanations

Monolog

The neuron fires on named entities and identifiers—that is, capitalized or multiword proper nouns, class/type names, and similar 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

ุง

-0.77

竜

-0.76

нда

-0.73

lod

-0.73

κης

-0.72

 разно

-0.71

dni

-0.70

liek

-0.69

 willing

-0.69

intas

-0.68

POSITIVE LOGITS

Mon

0.88

로운

0.83

 کافی

0.78

 perioden

0.77

Emin

0.77

 Жу

0.76

 هنر

0.75

 niggas

0.73

={}

0.73

íte

0.72

Activations Density 0.060%