INDEX

Explanations

years and dates

The neuron activates on numeric digit tokens (i.e. the individual characters of numbers, especially years).

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

 теперь

-1.45

atste

-1.42



-1.38

DebuggerNonUser

-1.37

Якщо

-1.34

 якщо

-1.34

Этот

-1.33

inverno

-1.32

។

-1.31

Jika

-1.30

POSITIVE LOGITS

1.63

1.39

1.36

1.33

1.29

1.26

 began

1.24

1.23

1.22

Activations Density 0.067%