INDEX

Explanations

star ratings

The neuron activates strongly on numeric tokens (digits or small numbers) wherever they appear.

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

):

-1.47

 Keeping

-1.41

can

-1.37

半年

-1.36

 October

-1.35

 outras

-1.34

 Before

-1.30

 Despite

-1.27

 other

-1.24

 Making

-1.23

POSITIVE LOGITS

ere

1.42

gruppe

1.34

1.30

1.18

 comrade

1.18

鐽

1.17

the

1.16

tion

1.13

یکن

1.12

 ウィメンズ

1.12

Activations Density 0.027%