INDEX

Explanations

jargon and specific terms

The neuron selectively fires on content-bearing (“open-class”) words—nouns, main verbs, adjectives, and adverbs—marking semantically significant tokens rather than high-frequency function words.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

ǘ

0.70

 laminar

0.60

 bureaucrats

0.60

 বিনামূল্যে

0.60

 multicolored

0.57

adiabatic

0.57

Multicolored

0.57

 abandonar

0.57

 крупных

0.56

 adiab

0.55

POSITIVE LOGITS

 Chiropractic

1.05

 Golf

0.92

 Hockey

0.92

 покер

0.89

Golf

0.89

 Equestrian

0.86

 golf

0.86

 paintball

0.86

Hockey

0.85

 hockey

0.84

Activations Density 1.049%