INDEX

Explanations

what is happening

The neuron fires on very common short function words (e.g. “to,” “an,” “say,” “means,” etc.).

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

 grossesse

-1.52

がありません

-1.46

 ваша

-1.26

 peligroso

-1.23

viamente

-1.21

 одним

-1.20

Disciplina

-1.20

することも

-1.19

ꜥ

-1.16

gebäude

-1.15

POSITIVE LOGITS

 most

2.05

 when

1.54

now

1.40

 some

1.38

 many

1.34

“

1.34

1.32

ire

1.30

 from

1.26

 they

1.25

Activations Density 0.132%