INDEX

Explanations

The neuron fires strongly on words naming a broad demographic group (e.g. “women”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tics

0.98

ларды

0.75

torrent

0.73

trades

0.72

tor

0.68

tien

0.66

ContentTypes

0.66

 hồn

0.66

yester

0.66

Noz

0.66

POSITIVE LOGITS

 étage

0.80

лага

0.79

м

0.79

 проекти

0.78

łach

0.78

Л

0.77

рд

0.74

গ্রেস

0.74

olaire

0.73

ف

0.73

Activations Density 0.000%