INDEX

Explanations

numbers with +/-

The neuron strongly activates on numeric tokens (e.g. standalone numbers or numeric measurements).

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

-1.61

])

-1.52

 faune

-1.40

˝

-1.36

りと

-1.34

];

-1.32

?..

-1.30

 bapt

-1.30

 watts

-1.29

POSITIVE LOGITS

 primeiras

1.53

of

1.50

 have

1.46

 surprising

1.45

 since

1.36

 xadrez

1.35

ﮨ

1.34

 depending

1.30

 interestingly

1.28

Установка

1.27

Activations Density 0.017%