INDEX

Explanations

Journal of, outtakes of, Yalla Shoot

The neuron detects multiword named‐entity or title phrases—especially capitalized titles of publications, products, shows, or place names.

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

by

-1.63

 that

-1.47

 with

-1.46

 there

-1.36

with

-1.27

蜥

-1.20

It

-1.20

what

-1.19

of

-1.16

POSITIVE LOGITS

Leírás

1.38

醅

1.34

embley

1.29

Rüyada

1.26

MultipartFile

1.26

奩

1.23

ăți

1.22

 farbe

1.21

 weiteres

1.20

 manej

1.20

Activations Density 0.378%