INDEX

Explanations

slight small little lower

The main thing this neuron does is detect words that express slightness or minor magnitude (e.g., “slight,” “slightest,” “change,” “small”).

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

UR

-1.47

-1.38

end

-1.36

 vogel

-1.36

-1.35

ven

-1.25

-1.23

œ

-1.22

FOUR

-1.21

 andré

-1.16

POSITIVE LOGITS

してしまった

1.37

 σήμερα

1.34

 variedades

1.34

 preferencias

1.32

wickshire

1.31

 dziś

1.28

 incomplète

1.28

it

1.27

灩

1.25

 persönlichen

1.24

Activations Density 0.067%