INDEX

Explanations

java.util.List

This neuron detects package‐path tokens in import or package declarations (e.g. Java/Android/SQLite package names).

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

̗

-1.27

if

-1.13

 until

-1.13

̻

-1.06

komme

-1.05

 bilhões

-1.04

 और

-1.04

寺の

-1.03

 conoció

-1.02

就可以了

-1.01

POSITIVE LOGITS

 трени

1.16

 espec

1.10

 estab

1.10

wrights

1.05

 fortal

1.02

 neces

0.99

 CAME

0.98

要点

0.97

摻

0.96

 extravag

0.94

Activations Density 0.007%