INDEX

Explanations

Swagger, cheese, Partnership

This neuron fires on the final multi‐letter fragments of longer compound words (often proper names or technical terms).

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

and

-1.84

it

-1.62

 effective

-1.51

勠

-1.42

or

-1.40

只

-1.39

时间

-1.39

phòng

-1.38

：(

-1.36

一

-1.36

POSITIVE LOGITS

meanor

1.66

 переписи

1.65

rawd

1.63

iyaki

1.63

 discern

1.62

 práci

1.59

ској

1.55

翃

1.55

うーん

1.53

various

1.52

Activations Density 0.113%