INDEX

Explanations

from one to another

This neuron specifically fires on code-related uppercase keywords and symbols (e.g. SQL/CTE keywords like CAST, CASE, JOIN, DECLARE or angle-bracket imports).

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

ſty

-1.51

 pág

-1.40

 Они

-1.39

 óptica

-1.38

 unwittingly

-1.36

 with

-1.30

 spolup

-1.30

 originalen

-1.30

 nabízí

-1.29

 unequivocally

-1.28

POSITIVE LOGITS

—

1.55

urations

1.38

1.30

 majestic

1.30

he

1.28

 gigantic

1.25

 delightful

1.23

1.22

},

1.21

 pero

1.18

Activations Density 0.169%