INDEX

Explanations

The neuron detects code blocks or structured text

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

0.37

 אם

0.33

 Waterloo

0.32

");

0.31

see

0.30

!”

0.30

0.29

not

0.29

!"

0.29

 thấy

0.29

POSITIVE LOGITS

لیفون

0.42

ائز

0.40

ām

0.39

melidir

0.39

īg

0.39

reactant

0.38

carbam

0.38

𝚓

0.38

ッケージ

0.38

ποί

0.36

Activations Density 0.000%

This feature has no known activations.