INDEX

Explanations

license headers and copyright notices

The neuron is sensitized to numeric literals (i.e. tokens consisting of numbers or number‐like strings).

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

᾽

-1.97

וּ

-1.88

ῤ

-1.70

ּוֹ

-1.53

to

-1.51

 continuación

-1.49

 தமி

-1.46

᾿

-1.44

 stück

-1.44

 frauen

-1.44

POSITIVE LOGITS

 овощи

1.87

When

1.76

шной

1.74

 иллюстрация

1.73

 Fähigkeit

1.73

Tujuan

1.72

当你

1.72

ὴν

1.72

While

1.71

During

1.69

Activations Density 0.002%