INDEX

Explanations

dates with bc or ad

Explanation of neuron 4 behavior: the main thing this neuron does is find numeric tokens—especially numbers in dates, years, centuries, and other measurements.

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

Wasn

-1.63



-1.54

 adesi

-1.53

髣

-1.49

Hopefully

-1.48

पुर

-1.43

 tři

-1.43

えない

-1.43

Staying

-1.42

äksi

-1.41

POSITIVE LOGITS

of

1.55

 endast

1.38

llten

1.30

 Тогда

1.27

 even

1.23

givers

1.22

 solely

1.21

municipi

1.20

 attempting

1.20

…」

1.18

Activations Density 0.044%