INDEX

Explanations

way into or through

This neuron detects numeric tokens (especially decimal numbers and large integers).

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

镲

-1.69

 usado

-1.65

饜

-1.64



-1.59

㤵

-1.48

𝙖

-1.47

 весь

-1.46

signia

-1.40

ωση

-1.38

 creada

-1.35

POSITIVE LOGITS

in

1.97

隸

1.55

</strong>

1.48

玙

1.42

 suggest

1.41

 Coloque

1.40

 appear

1.38

 many

1.36

]$}

1.34

呦

1.34

Activations Density 0.015%