INDEX

Explanations

a or an before words

The neuron activates on occurrences of the indefinite articles “a” or “an” (in any case).

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

 てき

-1.40



-1.38

también

-1.37

ebly

-1.30

 indirec

-1.28

 myös

-1.26

 olika

-1.25

 suppos

-1.24

uggest

-1.23

olving

-1.23

POSITIVE LOGITS

が

1.48

ENT

1.34

can

1.34

to

1.30

of

1.29

 schermata

1.28

 person

1.27

 could

1.25

や

1.17

AN

1.16

Activations Density 0.055%