INDEX

Explanations

accounts for percentage

The neuron fires on numeric tokens (counts, measurements, percentages, etc.).

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

-1.73

ها

-1.42

You

-1.41

年底

-1.39

derdag

-1.35

 тоже

-1.32

ين

-1.27

 samma

-1.25

värr

-1.20

也

-1.20

POSITIVE LOGITS

 than

1.68

of

1.55

heute

1.48

 gesamte

1.45

睒

1.34

都需要

1.30

 protože

1.30

 zusätz

1.30

 deoarece

1.29

 since

1.27

Activations Density 0.053%