INDEX

Explanations

download free content

The neuron detects section‐header and navigation labels related to press/media materials and downloads (e.g. “Press,” “Section,” “Downloads,” “Download,” “Kit”).

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

žena

-1.24

 tecnologias

-1.24

通过

-1.23

通過

-1.23

 кто

-1.22

 ayant

-1.20

哪怕

-1.17

neus

-1.15

who

-1.14

 tecnológica

-1.14

POSITIVE LOGITS

 think

1.21

 also

1.17

isters

1.14

њу

1.13

for

1.09

 acceptez

1.09

颗粒

1.09

 aimez

1.08

 загру

1.06

 perro

1.05

Activations Density 0.077%