INDEX

Explanations

beginnings of sentences and phrases

This neuron fires on tokens that occur in passages of direct speech (quoted dialogue).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

菫

-1.42

־ה

-1.39

ִּ

-1.38

 πάντα

-1.35

 бампер

-1.34

 suéter

-1.32

 menghubungi

-1.31

וּ

-1.31

 Gerechtigkeit

-1.31

Especies

-1.30

POSITIVE LOGITS

 your

2.20

you

1.77

my

1.65

 will

1.48

 have

1.23

by

1.22

our

1.20

._

1.16

하지

1.16

 चीज़

1.13

Activations Density 0.009%