INDEX

Explanations

evaluation following "it's"

The neuron fires on occurrences of the copular “be” (in its various forms: be, is, are, etc.) linking subjects and predicates.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

此外

0.86

<h2>

0.79

으로부터

0.75

Lorsque

0.74

Perhaps

0.73

Create

0.73

さまざまな

0.73

Asimismo

0.71

 vervolgens

0.71

Moreover

0.70

POSITIVE LOGITS

 gonna

1.59

 badass

1.55

 horrible

1.55

 crappy

1.51

 shitty

1.49

 kinda

1.48

 overrated

1.46

 messed

1.45

 weird

1.44

ってる

1.38

Activations Density 0.845%