INDEX

Explanations

academic context

This neuron fires strongly on multiword, capitalized scientific paper titles or citation headings.

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

 particularly

-1.83

 these

-1.70

 mariposas

-1.66

 poliuret

-1.65

 participado

-1.63

 velas

-1.62

 algas

-1.59

 silang

-1.57

 този

-1.55

 burbujas

-1.55

POSITIVE LOGITS

 emoji

1.59

 arXiv

1.58

 mesmo

1.55

https

1.50

 emojis

1.45

 microbiome

1.45

 https

1.44

 livestream

1.44

<u>

1.41

 instagram

1.41

Activations Density 0.327%