INDEX

Explanations

business and technical terms

The neuron fires on multisyllabic, morphologically complex content words—especially those built with common Latin- or Greek-derived affixes (-tion, ‑al, ‑ity/-ies, ‑ing, 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

or

-1.95

 when

-1.78

 because

-1.71

if

-1.56

 since

-1.54

 went

-1.49

can

-1.44

 they

-1.41

to

-1.38

 before

-1.36

POSITIVE LOGITS

 gants

1.38

 uren

1.26

benen

1.25

ERTY

1.23

 bouteilles

1.20

ysteme

1.17

 opinie

1.17

сных

1.15

ější

1.13

IONI

1.13

Activations Density 0.384%