INDEX

Explanations

I, my, am statements

This neuron fires on first‐person self‐references—instances where the author uses “I,” “I’m,” “my,” etc., to give personal opinions or describe their own experience.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 this

-1.77

这些

-1.65

 that

-1.56

鈎

-1.51

 BorderRadius

-1.33

この

-1.31

這些

-1.30

 همچنین

-1.30

bildet

-1.27

 after

-1.26

POSITIVE LOGITS

but

1.79

”.

1.45

 lorsqu

1.41

”).

1.34

σει

1.33

 lorsque

1.26

figured

1.24

 davor

1.24

).”

1.23

%).

1.22

Activations Density 0.125%