INDEX

Explanations

at a place

The neuron activates on standalone numeric tokens (e.g. years, ages, street numbers, section numbers).

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

 vetores

-2.00

 เรา

-1.84

籏

-1.73

殍

-1.73

ようになりました

-1.66

鹩

-1.63

 controladores

-1.62

でもある

-1.59

﹍﹍

-1.59

坌

-1.59

POSITIVE LOGITS

1.91

all

1.79

1.73

to

1.67

1.59

...

1.55

Perhaps

1.54

but

1.50

--

1.49

Activations Density 0.033%