INDEX

Explanations

you may not use or may occur

The neuron fires on formal disclaimer and licensing language—words used in product notices, warranty statements, and legal/license compliance text.

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

 suéter

-1.28

 blusa

-1.27

Là

-1.26

 plát

-1.26

鸲

-1.23

 uniformes

-1.20

 teşekkür

-1.20

 painel

-1.16

 esportes

-1.16

 intelig

-1.16

POSITIVE LOGITS

has

1.41

and

1.39

may

1.36

 will

1.30

can

1.29

𝑈

1.20

you

1.16

 which

1.13

of

1.09

，

1.09

Activations Density 0.001%