INDEX

Explanations

code components and structures

The neuron fires strongly on capitalized identifiers—especially PascalCase type, class, or component names (and related tokens) in import statements and declarations.

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

 will

-1.20

ところに

-1.16

any

-1.16

 terwijl

-1.05

 either

-1.05

ꗃ

-1.02

xxv

-1.01

にあります

-1.00

 precedenti

-1.00

POSITIVE LOGITS

<bos>

1.39

in

1.24

 вашего

1.13

 ваших

1.11

</h2>

1.05

лум

1.04

']))

1.04

潜力

1.03

參考文獻

1.01

 AppCompatTheme

1.00

Activations Density 0.021%