INDEX

Explanations

already previously was

The neuron fires on concrete, countable object nouns—things you can hold, give, or use (e.g. fish, pickles, toys, towels, cups, candy, 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

 these

-1.63

 each

-1.48

لكن

-1.36

هذه

-1.34

contrar

-1.30

 estadounid

-1.23

 évé

-1.21

鼕

-1.21

 your

-1.20

 there

-1.20

POSITIVE LOGITS

for

1.87

 already

1.75

 which

1.29

1.27

 уже

1.23

ALREADY

1.20

was

1.20

ссию

1.20

 previously

1.16

 już

1.16

Activations Density 0.058%