INDEX

Explanations

the followed by content type

The neuron detects imperative calls‐to‐action prompting the reader to access or consume content (e.g. “Listen to the podcast,” “Read the transcript,” “Read the Article”).

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

 gleichen

-1.55

to

-1.55

 genannten

-1.47

 höheren

-1.36

 ähnlichen

-1.34

 anklicken

-1.33

Қ

-1.27

 săptăm

-1.25

 gleiche

-1.24

 DIFFERENT

-1.24

POSITIVE LOGITS

 contagios

1.32

ífero

1.31

 comprador

1.26

设计的

1.26

 cocin

1.25

收购

1.25

 koning

1.23

潯

1.23

 mereka

1.22

 elefante

1.20

Activations Density 0.035%