INDEX

Explanations

you can see

This neuron fires on “as you can see”–style phrases used to point the reader’s attention to something.

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

-1.65

 trebui

-1.31

چگونه

-1.20

cyjnych

-1.16

 Honestly

-1.15

مقدمه

-1.12

 завжди

-1.12

Grath

-1.11

quelize

-1.10

してました

-1.09

POSITIVE LOGITS

1.70

of

1.34

 there

1.32

在你

1.30

we

1.20

ot

1.16

co

1.13

my

1.12

1.11

 here

1.10

Activations Density 0.019%