INDEX

Explanations

java imports and annotations

This neuron fires on code tokens that name types or attributes—particularly imported class or annotation identifiers.

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



-1.46

⦑

-1.43

 作り方

-1.42

Также

-1.30

Beskrivning

-1.30

 panik

-1.30

 sava

-1.29

 perut

-1.27

ACIONAL

-1.26

Zubereitung

-1.26

POSITIVE LOGITS

1.36

am

1.33

1.27

uk

1.23

1.16

1.15

)});

1.14

un

1.13

1.12

Activations Density 0.001%