INDEX

Explanations

almost everything, everywhere

The neuron responds to language about amounts, counts, measures, or time spans—i.e. quantitative or temporal references.

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

 trotz

0.76

 suggestive

0.74

 unorthodox

0.73

 mishap

0.66

 diminutive

0.66

 resilient

0.64

 pesar

0.64

 способствует

0.64

 pivoted

0.63

 misconception

0.63

POSITIVE LOGITS

すべての

1.66

 semua

1.48

 всіх

1.48

 everywhere

1.47

全ての

1.41

 جميع

1.40

 всички

1.40

 всех

1.39

 wszystkie

1.39

 모든

1.38

Activations Density 2.101%