INDEX

Explanations

distributed, license, category, Italy

The neuron predominantly activates on decimal numbers (floating‐point numeric tokens).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

-2.20

考えると

-1.48

 japonais

-1.45

 dibanding

-1.44

 inglese

-1.42

-1.34

They

-1.34

they

-1.30

 вашего

-1.29

ếp

-1.28

POSITIVE LOGITS

the

1.83

its

1.63

 favorece

1.47

 unterschied

1.45

zwischen

1.44

釭

1.44

珦

1.38

 trajo

1.34

 ilumina

1.33

 مجرد

1.33

Activations Density 0.008%