INDEX

Explanations

building components and issues

The neuron is strongly activated by words that appear in section headings or article titles (e.g. “Black mold,” “Floods,” “Blocked Drains,” “Moisture Vapor Barrier Kit,” 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

一丝

-1.48

already

-1.47

ggars

-1.39

出声

-1.34

 harris

-1.32

 écrire

-1.30

妩

-1.30

#########

-1.27

じっくり

-1.27

噎

-1.27

POSITIVE LOGITS

 your

2.03

our

1.58

 specchio

1.51

 zamiast

1.47

if

1.46

 cammin

1.41

してる

1.39

などで

1.37

 doppia

1.36

 votre

1.32

Activations Density 0.108%