INDEX

Explanations

code comments and programming structures

This neuron is triggered by long runs of whitespace (indented or blank lines).

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

 eenig

-1.42

 obligé

-1.41

 coût

-1.41

 payé

-1.41

 différents

-1.40

 acheté

-1.38

 właścic

-1.38

 épar

-1.37

 своих

-1.36

 véritable

-1.36

POSITIVE LOGITS

 because

1.54

 there

1.52

.”

1.45

1.30

</em>

1.27

there

1.26

<bos>

1.25

you

1.22

 during

1.20

轍

1.20

Activations Density 0.000%