INDEX

Explanations

eye

This neuron activates on the second part of “eye”-based compound phrases (e.g. “eye collection,” “eye study,” “eyes-free,” “eye-gasm”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

to

-1.47

禦

-1.42

 يول

-1.40

-1.38

hmigung

-1.37

 обеспечивает

-1.32

was

-1.30

-1.28

-1.27

 understand

-1.27

POSITIVE LOGITS

羮

1.64

sprzedam

1.55

llavero

1.51



1.46

 baumwolle

1.46

eye

1.46

夘

1.45

kaç

1.42

鱨

1.42

 eyes

1.41

Activations Density 0.039%