INDEX

Explanations

This neuron detects sequences of hexadecimal characters (e.g. GUIDs, secret keys) within identifiers.

Head long, spotlight on, pages

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(

-1.23

........

-1.18

EdgeInsets

-1.16

𝒙

-1.13

 Júnior

-1.13

本番

-1.13

 system

-1.12

 .....

-1.12

 while

-1.08

any

-1.06

POSITIVE LOGITS

 necessariamente

1.28

Expan

1.25

 frequentemente

1.22

ופת

1.21

licans

1.20

endroit

1.20

Citiți

1.19



1.19

 gewä

1.18

 aprobó

1.18

Activations Density 0.001%