INDEX

Explanations

citations and mathematical symbols

The neuron activates on tokens containing an asterisk (e.g. superscript “*” markers in formulas or 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

styleUrls

-1.73

殿下

-1.56

 botol

-1.54

 køb

-1.53

᩠

-1.52

kfree

-1.44

міні

-1.43

翆

-1.38

 rigtig

-1.38

lentes

-1.37

POSITIVE LOGITS

1.50

1.48

1.46

 néz

1.40

if

1.38

1.37

↵↵

1.31

 protože

1.31

 './../

1.30

1.27

Activations Density 0.018%