INDEX

Explanations

publishing press

The neuron fires on occurrences of publisher imprints labeled with “Press,” i.e. names of publishing houses ending in “Press.”

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

などなど

-1.71



-1.55

 hilfiger

-1.55

くちゃ

-1.55

öffnet

-1.51

禇

-1.49

的笑容

-1.48

 angelo

-1.45

 Again

-1.44

 می‌توان

-1.44

POSITIVE LOGITS

 will

1.44

derdag

1.39

 otras

1.38

妁

1.37

</strong>

1.35

";

1.33

 една

1.30

 nuestro

1.30

物

1.28

史上

1.27

Activations Density 0.011%