INDEX

Explanations

book, skates, 30

The neuron fires on product‐listing metadata headings and their high‐value descriptors (e.g. “Manufacturer:”, “Model:”, “Condition:”, “New”, “Used”, 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

Atentamente

-1.31

bungen

-1.24

さんへ

-1.22

eara

-1.21

 povin

-1.18

 year

-1.16

をご覧ください

-1.16

 each

-1.16

是什么意思

-1.15

お試し

-1.14

POSITIVE LOGITS

1.33

 saisons

1.19

zeß

1.16

dienne

1.12

많

1.12

 privé

1.11

 épo

1.07

都

1.05

tså

1.05

*}[

1.05

Activations Density 0.155%