INDEX

Explanations

car models and brands

The neuron responds strongly to tokens that are part of product or model identifiers (e.g. model names and version numbers).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits



-1.26

さてさて

-1.25

 marte

-1.25

 équipée

-1.24

 segredo

-1.23

 famoso

-1.21

⤷

-1.20

昨日は

-1.19

 semaines

-1.19

uye

-1.17

POSITIVE LOGITS

if

1.44

conos

1.27

致

1.25

ﾈ

1.23

izop

1.22

 anzunehmen

1.20

三种

1.18

 other

1.17

Filmography

1.16

<<"

1.15

Activations Density 0.065%