INDEX

Explanations

content labels

The neuron detects section headers or labels paired with item counts (e.g. “News 129,” “Articles 424,” “Posts 414,” etc.).

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

-1.42

 skru

-1.34

 veta

-1.31

-1.29

です

-1.28

Поэтому

-1.27

 kollu

-1.22

The

-1.22

 ESPECIAL

-1.22

メリカ

-1.20

POSITIVE LOGITS

1.68

are

1.65

Of

1.40

its

1.34

 must

1.30

 include

1.30

 kwamen

1.27

 lác

1.27

acimiento

1.26

1.25

Activations Density 0.112%