INDEX

Explanations

a writing instrument

The neuron is strongly activated by standalone numeric tokens and other number‐related data (e.g. statistics, dates, counts).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 saree

1.27

 Außen

1.21

 Kareena

1.20

☐

1.18

 quartzite

1.18

 heute

1.16

 mehndi

1.16

 Urdu

1.14

 supremum

1.13

 Mengen

1.13

POSITIVE LOGITS

In

1.67

sthe

1.53

:“

1.51

The

1.49

”“

1.48

the

1.45

and

1.41

është

1.41

There

1.40

ம்த

1.39

Activations Density 2.262%