INDEX

Explanations

dates and times

The neuron is primarily activated by numeric tokens (e.g. standalone digits or numbers, often in date or quantitative contexts).

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

pady

-1.45

 cependant

-1.38

 чёрный

-1.37

-1.36

de

-1.34

了些

-1.32

we

-1.30

its

-1.30

がついて

-1.30

mu

-1.26

POSITIVE LOGITS

↵↵

2.39

 rantai

1.60

sió

1.45

렛

1.37

已經

1.34

現在

1.32

 menangis

1.31

 memutus

1.30

 that

1.30

łości

1.30

Activations Density 0.033%