INDEX

Explanations

date and city

The neuron strongly activates on numeric tokens (years, dates, times, and other 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

cellor

-1.59

Apesar

-1.48

CLUDING

-1.46

Após

-1.39

entary

-1.38

enablog

-1.36

successfully

-1.35

᯾

-1.34

 χρησιμοποι

-1.34

 Majority

-1.33

POSITIVE LOGITS

1.80

1.45

↵

1.45

をする

1.37

1.34

それを

1.33

我们的

1.25

着他

1.23

 живота

1.21

我们

1.20

Activations Density 0.232%