INDEX

Explanations

The neuron detects promotional or sales‐pitch language describing products or services (i.e., advertisement text).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 three

0.43

 reported

0.41

two

0.41

 putative

0.41

 drie

0.39

van

0.39

0.38

da

0.38

 only

0.36

 stated

0.36

POSITIVE LOGITS

érm

0.52

qualiter

0.52

éseket

0.51

avasena

0.50

ümler

0.49

ătur

0.49

óticos

0.49

ađ

0.48

 Ана

0.48

रिडोर

0.48

Activations Density 5.616%