INDEX

Explanations

what you should do next

The neuron flags content-bearing words—mostly nouns and verbs—rather than grammatical or function words.

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

 באמצעות

0.63

(),

0.61

 pursuant

0.59

’,

0.58

)(

0.57

 của

0.57

(,

0.57

()

0.54

 включают

0.54

POSITIVE LOGITS

ってる

0.79

 আমি

0.74

俺

0.73

ってます

0.70

আমি

0.68

但是我

0.68

 Leute

0.67

我的

0.66

 আমার

0.65

 meinem

0.64

Activations Density 0.001%