INDEX

Explanations

categories of services or entities

The neuron spotlights salient content nouns—especially in compound terms or section titles—highlighting the main topical or proper‐name words in a passage.

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

 that

-2.05

}();

-1.83

."/

-1.68

}());

-1.47

//});

-1.46

():

-1.42

atação

-1.42

tling

-1.38

 actrice

-1.37

.='

-1.37

POSITIVE LOGITS

鱟

1.48

 erad

1.23

歯

1.20

 spearheaded

1.20

the

1.18

gorra

1.18

 seiz

1.17

 disciplined

1.16

イエロー

1.16

 vijf

1.16

Activations Density 3.747%