INDEX

Explanations

modelmodel numbers or series

The neuron strongly responds to numeric tokens—particularly individual digits or long digit‐only sequences in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

⦊

-1.80

</strong>

-1.74

-1.64

する

-1.53

真的很

-1.50

Cliquez

-1.50

-1.48

熜

-1.48

终于

-1.46

ELTS

-1.45

POSITIVE LOGITS

rijven

1.51

と呼ば

1.45

 есть

1.42

 construite

1.40

 tòa

1.38

 Flowers

1.37

 длина

1.37

 дети

1.34

stawa

1.28

ll

1.27

Activations Density 0.016%