INDEX

Explanations

operators and assignments

The neuron fires on numeric literals—especially floating‐point constants—embedded in code.

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

it

-1.38

we

-1.27

if

-1.23

is

-1.22

–

-1.05

 които

-1.05

 обычно

-1.02

catalogo

-1.01

of

-1.01

POSITIVE LOGITS

</h4>

1.37

 LIABLE

1.32

 navide

1.29

 verzi

1.28

 diventato

1.27

⿷

1.20

、“

1.19

 personale

1.19

 époque

1.16

 rythme

1.15

Activations Density 0.016%