INDEX

Explanations

This neuron detects closing delimiters in structured markup—such as the end of JSON objects (“},”), XML/HTML tags (“>”), and Markdown links or parentheses.

unauthorized activity detected

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

for

-1.55

in

-1.41

-1.30

GEBRA

-1.29

 Mentre

-1.15

 their

-1.14

-1.12

or

-1.10

you

-1.10

he

-1.08

POSITIVE LOGITS

↵↵

1.50

 protegida

1.32

 indicado

1.23

 desactivar

1.20

 erschienen

1.19

 preguntar

1.16

 suministros

1.15

 rilass

1.14

 specifiche

1.13

 instalada

1.13

Activations Density 0.001%