INDEX

Explanations

free from trouble or burden

The neuron detects words formed with suffixes that denote absence, protection, or ease—e.g. “-proof,” “-free,” “-safe,” “-resistant,” and “-effective.”

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

to

-3.55

for

-2.98

in

-2.75

on

-2.38

of

-2.33

 that

-2.17

at

-2.16

by

-2.06

 from

-1.89

 beragam

-1.88

POSITIVE LOGITS

Недо

1.65

 prochaines

1.50

 हम

1.45

我們

1.45

 actuelle

1.41

 ytterligare

1.40

forgettable

1.39

 первом

1.39

 シール

1.37

 Skyscanner

1.37

Activations Density 0.164%