INDEX

Explanations

thinking and believing

The neuron fires on first-person opinion or hedge phrases—e.g. “I think” (and its equivalents like “penso che”).

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

婼

-1.58

<tfoot>

-1.52

 considérons

-1.52



-1.52

))^

-1.45

</tfoot>

-1.44

えます

-1.41

entlichen

-1.41

are

-1.40

諌

-1.40

POSITIVE LOGITS

1.44

在大

1.43

 dañ

1.39

步行

1.38

 vermag

1.36

 hvis

1.34

</sup>

1.34

 persönlichen

1.31

 muerta

1.29

ちょっと

1.28

Activations Density 0.011%