INDEX

Explanations

dates and months

The neuron activates on numeric tokens representing calendar years (four-digit date mentions).

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

鏷

-1.80

</th>

-1.61

为什么会

-1.52



-1.52

richts

-1.51

 mentales

-1.42

waniu

-1.42

 sebagian

-1.42

 Democrá

-1.38

肫

-1.38

POSITIVE LOGITS

of

1.57

↵

1.30

 другом

1.28

がこの

1.27

 demselben

1.23

 jenem

1.22

 months

1.21

的で

1.20

 mieście

1.19

Activations Density 0.033%