INDEX

Explanations

Microsoft Word, Illustrator, Access

The main thing this neuron does is detect mentions of Microsoft Office (and related) application names.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

in

-1.68

‘

-1.58

 onPostExecute

-1.56

 Tahun

-1.41

 hver

-1.40

 спор

-1.39

—

-1.38

 This

-1.37

 that

-1.36

 dasar

-1.34

POSITIVE LOGITS

呖

1.73

 estadía

1.66

饜

1.64

 automó

1.55

 câmbio

1.49

帔

1.49

 dança

1.48

還會

1.47

 trabalhar

1.46

 senhora

1.45

Activations Density 0.056%