INDEX

Explanations

links and urls

The neuron fires on tokens that are part of structured numeric elements—e.g. digits in section numbers, file paths, version numbers, citations, or other numeric identifiers.

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

</strong>

-3.72

</i>

-3.33

</b>

-3.16

</em>

-3.02

</u>

-2.27

</code>

-1.58

-1.51

</h2>

-1.48

–

-1.41

--

-1.34

POSITIVE LOGITS

 Diſ

1.59

↵↵↵

1.34

 desbloquear

1.26

 новый

1.26

;\\

1.23

 cimetière

1.23

 comunicaciones

1.21

 Geräten

1.20

 cabelo

1.20

 всегда

1.19

Activations Density 0.378%