INDEX

Explanations

chemical notation or code elements

The neuron strongly activates on named references to formal institutions or government bodies (e.g. “Commission,” “Government,” “Pakistan,” “Afghanistan”).

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

vis

0.89

0.78

grids

0.73

 그러니까

0.72

eu

0.69

cott

0.68

roads

0.68

greet

0.68

amp

0.66

endous

0.66

POSITIVE LOGITS

 принимает

0.93

Ссы

0.86

 остальные

0.82

 девушки

0.82

 сначала

0.82

 ต้น

0.82

 parthenogenetic

0.82

 иллюстра

0.81

 владель

0.80

 воздей

0.80

Activations Density 0.000%