INDEX

Explanations

and whatnot

The neuron spikes on the head word of an “X of Y” nominal phrase (i.e. nouns immediately followed by the preposition “of”).

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

are

-1.79

 always

-1.55

 will

-1.48

There

-1.48

If

-1.43

 were

-1.42

เช่น

-1.34

 such

-1.34

Those

-1.34

by

-1.33

POSITIVE LOGITS

язы

1.43

èm

1.42

 antiguo

1.38

 می‌گو

1.36

 mystérie

1.34

ιος

1.34

ཫ

1.33

 historias

1.30

Aran

1.30

сви

1.29

Activations Density 0.285%