INDEX

Explanations

receive updates via email

The main thing this neuron does is detect newsletter or email‐subscription call-to-action phrases (e.g. “Receive the latest … in your inbox”).

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

 dokonce

-1.55

 because

-1.39

 beeind

-1.34

いかがでしたか

-1.28

 tiež

-1.27

 cómoda

-1.27

ポイントは

-1.26

 actually

-1.24

 totiž

-1.24

 sanitarias

-1.22

POSITIVE LOGITS

 ۱۱

1.59

 ۱۷

1.57

 ۱۶

1.54

 ۱۸

1.51

 NSCoder

1.45

ícias

1.41

corrhi

1.40

<td>

1.39

cyklopedia

1.38

 ۱۴

1.35

Activations Density 0.049%