INDEX

Explanations

following 'not'

The neuron flags negation words (forms of “not,” “never,” “n’t,” 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

any

-1.63

 until

-1.38

してしまう

-1.25

很喜欢

-1.22

may

-1.22

 there

-1.20

but

-1.20

 kuten

-1.16

しているので

-1.13

 anything

-1.11

POSITIVE LOGITS

 also

2.19

 almeno

1.52

 também

1.50

 также

1.38

 myös

1.35

 already

1.34

 también

1.34

 atleast

1.33

also

1.23

 également

1.23

Activations Density 0.021%