INDEX

Explanations

DIY projects make

The neuron detects document section headings or titles—especially DIY/tutorial headers (e.g. “DIY: …,” all-caps or prominent title lines).

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

ִּ

-1.63

ַּ

-1.53

ּוֹ

-1.45

ֶּ

-1.36

ְּ

-1.27

ֵּ

-1.27

 الأولى

-1.26

 أيض

-1.24

 бампер

-1.20

 inalámbrica

-1.16

POSITIVE LOGITS

ὕ

1.81

you

1.69

ׇ

1.48

DIY

1.34

 tutorial

1.30

 project

1.30

cán

1.24

onekana

1.21

 χρησιμοποι

1.20

 الجلد

1.20

Activations Density 0.069%