INDEX

Explanations

compound noun phrases

The neuron fires on isolated technical‐ or domain‐specific jargon nouns (e.g. “property,” “tax,” “bar,” “hand,” “alarm,” “pattern,” “hairpin,” “smooth,” “brady,” “signals,” “signs”).

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

—

-1.39

-1.27

-1.13

Obrázky

-1.13

という感じ

-1.13

があります

-1.12

).

-1.09

種類

-1.05

POSITIVE LOGITS

and

1.86

に似

1.45

が好き

1.41

 hvilken

1.24

pisode

1.20

他和

1.20

 revelado

1.19

⤹

1.18

 hvilke

1.16

AutoModerator

1.15

Activations Density 0.050%