INDEX

Explanations

automobile makes and models

The neuron selectively responds to numeric tokens—i.e. any standalone sequence of digits (model numbers, years, measurements, 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

 like

-1.48

 such

-1.37

-1.28

 next

-1.28

 einzelne

-1.25

 including

-1.22

 number

-1.21

…

-1.21

ed

-1.20

 Their

-1.20

POSITIVE LOGITS

findung

1.55

onaise

1.41

vlád

1.40

ograft

1.39

淠

1.38

gradable

1.38

мії

1.36

鏄

1.35

 géant

1.34

 derfor

1.34

Activations Density 0.028%