INDEX

Explanations

code comments and headers

The neuron fires on numeric tokens (literal numbers) wherever they appear in code or comments.

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

mengg

-1.38

versen

-1.36

aktionen

-1.34

куртка

-1.32

sisting

-1.30

ASKET

-1.27

 Wohnung

-1.27

‌

-1.26

dengan

-1.25

menjadi

-1.25

POSITIVE LOGITS

is

1.81

may

1.80

they

1.42

 tendrá

1.41

atasaray

1.29

EOUS

1.25

valla

1.25

 yourself

1.25

 there

1.24

 mungkin

1.23

Activations Density 0.004%