INDEX

Explanations

code comments and definitions

The neuron fires on C/C++ preprocessor macro definitions (e.g. lines with “#define” or the word “definition”).

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

 задний

-1.65

 передний

-1.55

 трубы

-1.49

 bemerk

-1.40

ׇ

-1.38

 katholischen

-1.37

を使って

-1.34

 Könige

-1.33

 poderoso

-1.32

hadores

-1.29

POSITIVE LOGITS

at

1.61

ּוֹ

1.45

ְּ

1.40

コミック

1.36

 will

1.36

 парфю

1.35

謁

1.34

 خریدار

1.32

ִּ

1.31

1.30

Activations Density 0.026%