INDEX

Explanations

every aspect

The neuron detects intensive “every” constructions that express totality or absolute scope (e.g. “in every way,” “in every sense,” “by every measure,” “every possible combination”).

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

vaných

-1.59

 volando

-1.57

Creates

-1.49

 mirando

-1.45

kwür

-1.45

 排骨

-1.45

めない

-1.43

 公仔

-1.43

yf

-1.43

 estadounid

-1.41

POSITIVE LOGITS

1.84

1.76

'';

1.70

1.67

1.63

1.61

 verbess

1.52

1.51

я

1.48

Activations Density 0.012%