INDEX

Explanations

The neuron fires strongly on big, hype-style adjectives and buzzwords (e.g. “compelling,” “effective,” “motivating,” “new”) used in promotional or corporate-speak.

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

()

0.33

The

0.28

the

0.27

0.26

 के

0.26

〉

0.25

}}{\

0.25

dhe

0.25

of

0.25

POSITIVE LOGITS

 impactful

0.48

 personable

0.46

 unassuming

0.46

 multifaceted

0.46

 lucrative

0.46

 empathetic

0.44

 soulful

0.43

 eclectic

0.43

 adaptable

0.43

 oppressive

0.43

Activations Density 1.262%