INDEX

Explanations

carried out

The neuron selectively activates on three-digit numeric 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

</h1>

-2.58

</b>

-1.95

᯾

-1.90

ೞ

-1.88

of

-1.81

恭敬

-1.80

StoryboardSegue

-1.76

↵

-1.73

巿

-1.72

 ระ

-1.66

POSITIVE LOGITS

2.59

an

2.52

榦

2.42

</em>

2.36

祤

2.19

on

2.14

才

2.13

僂

2.11

 vantagens

2.08

 ajouté

2.00

Activations Density 0.006%