INDEX

Explanations

names and foreign words

The neuron fires on tokens that are parts of proper nouns—especially geographic place names.

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

・「

-2.08

-2.06

门外

-1.87

when

-1.86

结论

-1.84

鯷

-1.83

本日は

-1.70

こんばん

-1.68

さてさて

-1.66

 потім

-1.64

POSITIVE LOGITS

,"

2.31

拗

1.91

了她

1.83

 praktisch

1.80

?"

1.77

咄

1.77

'',

1.76

 travaillent

1.74

è

1.73

 refugiados

1.73

Activations Density 0.080%