INDEX

Explanations

import statements

This neuron fires on salient, topic-specific content words—particularly proper names, technical identifiers, code tokens, and other uncommon nouns.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

как

0.63

此外

0.61

；

0.61

Ра

0.60

.}}

0.57

"},

0.56

}\

0.56

}.

0.55

).}

0.55

`}

0.55

POSITIVE LOGITS

 import

0.89

 olduğunu

0.83

 said

0.80

َالَ

0.79

 gerektiğini

0.75

 dijo

0.72

 olmadığını

0.69

 einen

0.68

 самую

0.68

czną

0.65

Activations Density 0.001%