INDEX

Explanations

assert statements

The neuron chiefly fires on numeric constants and compile‐time constant tokens in code (e.g. literal numbers, macro or template constants).

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

or

-1.54

for

-1.52

 они

-1.51

 believes

-1.51

 these

-1.43

 indicato

-1.42

 seems

-1.42

 While

-1.40

at

-1.39

 Often

-1.37

POSITIVE LOGITS

ִּ

1.47

めちゃくちゃ

1.44

և

1.39

 EARLY

1.34

 présenter

1.30

言った

1.28

뚠

1.28



1.28

siasme

1.28

 dilapidated

1.26

Activations Density 0.010%