INDEX

Explanations

forum post counts and signatures

The neuron specifically detects forum-style user profile metadata (e.g. “Joined,” “Posts,” colons and their accompanying numbers).

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

to

-1.90

</h2>

-1.52

as

-1.48

an

-1.44

 (“

-1.40

-1.30

for

-1.29

蹊

-1.24

“

-1.21

'])){

-1.16

POSITIVE LOGITS

 lentamente

1.44

 avanzado

1.38

dú

1.37

 obviamente

1.37

 первое

1.36

 muertes

1.34

 acolo

1.34

 afectado

1.34

 dismin

1.33

 reempla

1.30

Activations Density 0.027%