INDEX

Explanations

traveling, gold, hello, over

The neuron never fires—it’s essentially inactive and doesn’t detect any pattern.

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

ができる

-1.23

謾

-1.16

लों

-1.14

/**

-1.08

譏

-1.08

quelize

-1.05

 mater

-1.05

utando

-1.03

 ſeveral

-1.03

служба

-1.02

POSITIVE LOGITS

<strong>

1.50

//

1.27

 trovano

1.21

’

1.13

 créées

1.12

当

1.10

↵↵↵

1.09

!"

1.07

Activations Density 0.004%