INDEX

Explanations

what iswhy is

This neuron detects section or article headings and other title-like tokens (e.g. numbered list headings, “What…,” “How to…,” “Understanding…,” 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

 такъ

-1.25

 този

-1.20

 célè

-1.17

 новом

-1.17

 тази

-1.16

 hvit

-1.16

 възможно

-1.15

 държа

-1.13

 blauw

-1.13

 всичко

-1.11

POSITIVE LOGITS

isn

1.52

</h2>

1.30

</h3>

1.28

 your

1.24

</strong>

1.20

</h4>

1.18

 doesn

1.17

</em>

1.15

1.10

 aren

1.05

Activations Density 0.006%