INDEX

Explanations

code snippets in multiple languages

The neuron specializes in detecting uppercase code identifiers and macro constant fragments.

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

guigu

-1.38

grats

-1.27

 aveo

-1.24

julie

-1.23

levis

-1.23

 this

-1.20

jaket

-1.19

nivea

-1.16

 Paesi

-1.16

uxx

-1.15

POSITIVE LOGITS

brtc

1.46

 рецепта

1.28

autres

1.28

ebly

1.27

فوائد

1.23

Folge

1.21

Zutaten

1.20

alians

1.17

 njego

1.17

still

1.16

Activations Density 0.038%