INDEX

Explanations

determiner followed by noun

The neuron is detecting common English function words—especially articles, pronouns, and prepositions (e.g. “the,” “in,” “on,” “for,” “my,” “their,” etc.).

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

 его

-1.76

its

-1.55

-1.50

 coû

-1.41

 lika

-1.38

 wert

-1.38

-“

-1.34

 (“

-1.33

 nutz

-1.31

 Qaeda

-1.31

POSITIVE LOGITS

1.59

 olika

1.46

or

1.44

zei

1.41

خدمات

1.39

1.38

 résultat

1.37

 různ

1.36

1.34

Vente

1.34

Activations Density 0.211%