INDEX

Explanations

multilingual enumerations and lists

The neuron strongly activates on very frequent, common English words (e.g. “the,” “of,” “world,” “like,” “typical,” “your,” “can”).

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

his

-1.55

STRUCTIONS

-1.33

 keputusan

-1.32

...');

-1.27

 diferença

-1.26

 kemajuan

-1.23

 perquè

-1.23

yaszt

-1.20

 desempe

-1.20

 kepentingan

-1.16

POSITIVE LOGITS

to

1.45

这里

1.34

不管是

1.32

Бра

1.29

无论是

1.23

많

1.23

❅

1.22

 olacaktır

1.20

Here

1.16

Zubereitung

1.14

Activations Density 0.027%