INDEX

Explanations

multi-lingual specific names

The neuron primarily activates on numeric tokens (e.g. track lengths, chart positions, years, counts) in the text.

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

of

-1.71

our

-1.62

-1.48

the

-1.28

 that

-1.27

-1.24

 their

-1.17

 into

-1.15

-1.13

POSITIVE LOGITS

 många

1.43

 景色

1.35

 möjlighet

1.31

Karena

1.23

hoku

1.20

Luxem

1.19

tså

1.16

伟大

1.13

Werbung

1.13

Nederland

1.11

Activations Density 0.049%