INDEX

Explanations

foreign words and prepositions

This neuron is highly sensitive to common English function words (e.g. short, high‐frequency prepositions and conjunctions like “by,” “and,” and “of”).

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

of

-2.88

in

-1.99

to

-1.91

was

-1.72

on

-1.55

があります

-1.52

 have

-1.48

but

-1.48

鴣

-1.45

at

-1.44

POSITIVE LOGITS

’

2.16

 rafra

1.69

unser

1.59

 chaqueta

1.58

 capucha

1.58

1.55

 fraî

1.55

<bos>

1.54

 sudadera

1.53

 gorra

1.52

Activations Density 0.043%