INDEX

Explanations

variable declarations and access modifiers

The neuron strongly activates on the keyword “private” (i.e. private visibility modifiers).

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

ąp

-1.53



-1.52

 even

-1.48

if

-1.43

 personlig

-1.40

 detalj

-1.38

 bilde

-1.38

 arbeid

-1.36

 involucrados

-1.35

 capucha

-1.34

POSITIVE LOGITS

”,

1.63

and

1.52

on

1.52

The

1.45

</h2>

1.37

at

1.30

udahkan

1.25

苅

1.22

…”

1.22

but

1.21

Activations Density 0.001%