INDEX

Explanations

passive constructions or statements about agents

We are, We have

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 belonged

0.47

 teníamos

0.46

 hadden

0.44

 tenían

0.43

真的很

0.41

 находиться

0.41

belong

0.40

 existed

0.40

 pertenecen

0.39

 avevano

0.39

POSITIVE LOGITS

被

1.63

被

1.62

ถูก

1.57

也被

1.52

 ถูก

1.47

就被

1.47

都被

1.46

会被

1.41

而被

1.29

被人

1.26

Activations Density 0.161%