INDEX

Explanations

let it be

The neuron fires on imperative “let … be” or similar “let X Y” constructions (e.g. “let it be,” “let there be chaos,” “let this be true”).

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

 huellas

-1.90

 medier

-1.87

 netts

-1.79

 medarbe

-1.73

 arkite

-1.72

 medlems

-1.71

 để

-1.71

%),

-1.70

uscular

-1.67

缑

-1.66

POSITIVE LOGITS

 Itália

1.70

 プレス

1.66

 When

1.57

шей

1.56

It

1.48

 There

1.48

茄子

1.46

 Unlike

1.44

 Когда

1.42

 Jér

1.42

Activations Density 0.010%