INDEX

Explanations

past years, decades, or a period

The neuron detects tokens that mark temporal references (e.g. words meaning “past,” “previous,” “next,” or similar time indicators) across different languages.

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

釭

-1.88

from

-1.87

-1.85

嶮

-1.85

März

-1.85

慫

-1.84

僕は

-1.82

 städ

-1.78

往往

-1.76

étoit

-1.76

POSITIVE LOGITS

縯

2.39



2.02

ൎ

2.00

 escova

1.96

Our

1.96

ครื่อง

1.95

 alça

1.94

兎

1.92

 Beschreibung

1.91

 toalha

1.88

Activations Density 0.004%