INDEX

Explanations

toward a model

The neuron specifically activates on the directional preposition “toward” (or “towards”) and its immediately following article (e.g. “a” or “an”) in section titles or headings.

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

 aqueles

-1.68

 aktiviteter

-1.65

 speelt

-1.55

非常的

-1.55

潾

-1.54

 monedas

-1.49

ーーー

-1.48

に入り

-1.48

 tornillos

-1.48

 liknande

-1.47

POSITIVE LOGITS

1.66

so

1.45

</em>

1.34

as

1.30

万里

1.29

 ideas

1.28

an

1.24

 escritório

1.22

."

1.21

.”

1.19

Activations Density 0.055%