INDEX

Explanations

copyright notices

The neuron fires on numeric literals containing a decimal point (i.e. floating‐point 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

Also

-2.38

After

-1.92

If

-1.88

Those

-1.80

Before

-1.78

 also

-1.77

mathrm

-1.73

 unusually

-1.72

People

-1.69

Usually

-1.69

POSITIVE LOGITS

 باسلام

1.76

 praticamente

1.65

 Dinas

1.65

 technisch

1.62

 zoude

1.61

螃

1.60

"",

1.58

cising

1.55

 frequentemente

1.54

 breathtaking

1.52

Activations Density 0.003%