INDEX

Explanations

one of

This neuron fires on English indefinite or quantitative determiners and pronouns—especially “one” (and related words like “some,” “that,” “first,” etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

but

-1.41

 usine

-1.36

by

-1.34

 have

-1.26

どちらも

-1.22

xxi

-1.20

ANSA

-1.18

เดียว

-1.16

 Sometimes

-1.15

don

-1.14

POSITIVE LOGITS

 that

2.31

 they

1.55

ที่

1.52

 them

1.34

bluza

1.30

 zacz

1.28

 ones

1.27

 budú

1.27

__':

1.27

 medži

1.26

Activations Density 0.037%