INDEX

Explanations

why, how, was questions

The neuron strongly activates on interrogative cues—question words (what, if, could, why, etc.) and question-mark punctuation signaling a question.

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

we

-1.52

-1.45

et

-1.44

這些

-1.36

my

-1.34

其余

-1.28

our

-1.26

 katk

-1.26

 your

-1.22

 также

-1.21

POSITIVE LOGITS

 something

1.41

batasan

1.35

corrência

1.34

himself

1.33

tues

1.31

ætte

1.25

drá

1.22

antula

1.22

ต้องการ

1.20

捫

1.20

Activations Density 0.013%