INDEX

Explanations

Official job titles and roles

The neuron fires on words indicating official designations or titles (e.g., “designated,” “hitter,” “officer,” “ombudsman”).

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

 verzi

-1.53

 immense

-1.49

縢

-1.45

 harten

-1.44

色々と

-1.43

 emblematic

-1.41

 間取り

-1.39

 astounding

-1.39

意外と

-1.38

 torba

-1.38

POSITIVE LOGITS

for

2.06

 there

2.05

at

1.80

 told

1.73

if

1.72

 after

1.71

 just

1.59

 There

1.57

in

1.55

 will

1.52

Activations Density 0.075%