INDEX

Explanations

java/android package imports

The neuron detects import statements (specifically the keyword “import” at the start of Java package imports).

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

AdvEx

-1.38

 nachher

-1.30

分かります

-1.09

 einzeln

-1.06

 тө

-1.05

🛻

-1.04

this

-1.03

眺

-1.02

見ても

-1.01

many

-1.00

POSITIVE LOGITS

ใ

1.17

 οὐ

1.16

त्र

1.14

<blockquote>

1.10

ȃ

1.06

$\$

1.05

 fosfor

1.05

Becky

1.02

いかがでしたか

1.01

 ноября

1.00

Activations Density 0.008%