INDEX

Explanations

punctuation and numbers

The neuron fires on structural or metadata markers and numeric/citation tokens—things like slashes, equals signs, equation/case labels, pipes, and other formatting punctuation used in headings, citations, or document metadata.

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

on

-1.64

 will

-1.50

 work

-1.43

can

-1.40

 allows

-1.30

 huge

-1.29

 have

-1.26

in

-1.25

 different

-1.23

 makes

-1.23

POSITIVE LOGITS

 solides

1.59

 nacht

1.47

 schild

1.44

 versie

1.44

 vendus

1.43



1.40

 maniere

1.38

 stoppen

1.37

 serien

1.30

 ſmall

1.30

Activations Density 0.117%