INDEX

Explanations

wrapping materials and actions

The neuron fires on words related to wrapping or covering materials/actions (e.g. wrap, wrapped, plastic wrap, bubble wrap).

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

 dış

-1.38

bereiche

-1.36

WTA

-1.35

፫

-1.33

 grise

-1.30

几乎

-1.30

doo

-1.27

𝗶

-1.24

izawa

-1.24

EMC

-1.24

POSITIVE LOGITS

not

1.66

 been

1.47

もしくは

1.45

 nemlig

1.37

 ہم

1.35

wrapping

1.31

 previsto

1.30

 когато

1.27

we

1.27

または

1.23

Activations Density 0.029%