INDEX

Explanations

code syntax and programming terms

The neuron fires on unusually long, multiword code tokens or identifiers (e.g. long variable names, function/endpoint strings, or constant names).

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

 this

-2.48

of

-2.20

 even

-1.56

our

-1.49

my

-1.42

 these

-1.42

how

-1.36

 such

-1.35

 that

-1.34

 have

-1.32

POSITIVE LOGITS

 وما

1.50

'))

1.46

”

1.39

 همان

1.38

 vuelve

1.34

锲

1.34

aturi

1.31

 novamente

1.31

сколь

1.30

водится

1.30

Activations Density 0.035%