INDEX

Explanations

code class keyword

The neuron activates on Java class literal references (i.e. occurrences of “.class”).

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

of

-3.22

and

-1.48

\&

-1.44

籥

-1.40

 this

-1.40

ету

-1.38

for

-1.36

好不容易

-1.34

жат

-1.32

 tireless

-1.30

POSITIVE LOGITS

1.77

 américains

1.63

However

1.61

1.59

 الكهرب

1.50

的问道

1.48

 gjort

1.46

1.45

 rotación

1.45

 замети

1.45

Activations Density 0.001%