INDEX

Explanations

cause of or causes of

This neuron detects mentions of “cause/causes” (discussion of what causes something).

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

…」

-1.88

could

-1.76

the

-1.76

would

-1.73

?」

-1.72

橚

-1.72

Compared

-1.68

又是一

-1.67

ी

-1.65

 店内

-1.64

POSITIVE LOGITS

↵↵

1.95

↵

1.91

in

1.63

册

1.53

of

1.46

 también

1.42

su

1.39

かず

1.38

кость

1.38

 минуту

1.37

Activations Density 0.043%