INDEX

Explanations

8 followed by symbols

The neuron strongly activates on numeric tokens (e.g. phone numbers, measurements, percentages, dates, prices, etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 Mekong

1.45

 Curator

1.41

тбол

1.38

1.33

 Faculties

1.30

 outbursts

1.29

 Responsible

1.26

ز

1.26

ocks

1.24

boy

1.23

POSITIVE LOGITS

ل

2.45

ка

2.27

氇

1.96

ку

1.79

 gigi

1.70

ieth

1.61

لل

1.55

 povas

1.54

لای

1.52

ن

1.52

Activations Density 6.320%