INDEX

Explanations

start of new sentences/items

The neuron responds strongly to passages that enumerate a subject’s attributes or components (i.e. feature-oriented lists).

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

 если

-1.37

if

-1.36

">(</

-1.31

 antworte

-1.30

its

-1.27

=)

-1.26

'],

-1.26

 özellikle

-1.25

))))))))

-1.24

！"

-1.21

POSITIVE LOGITS

1.48

 όλα

1.28

踯

1.23

 исчез

1.20

車が

1.20

 farmaco

1.19

etc

1.19

 刺繍

1.18

わかった

1.18

 Spotted

1.17

Activations Density 0.091%