INDEX

Explanations

factual statements

This neuron detects formal section-heading or document-structure tokens (e.g. “Field of the Invention,” “Court of Appeals,” “Abstract,” “Learn more about,” 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

 then

-1.51

to

-1.34

 That

-1.25

再加上

-1.16

 Then

-1.16

 tangled

-1.15

 about

-1.13

讪

-1.11

くださった

-1.09

POSITIVE LOGITS

 your

1.57

();

1.34

 will

1.27

your

1.23

 pierna

1.22

 bluzka

1.21

 kvě

1.20

ᯅ

1.20

kese

1.20

 paño

1.18

Activations Density 0.501%