INDEX

Explanations

this. or self

The neuron detects object‐context references in code, i.e. occurrences of “this” or “self” (often followed by a dot or underscore) indicating property or method access.

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

-1.67

-1.59

 According

-1.42

中秋

-1.39

 Those

-1.39

 ПОД

-1.38

先日

-1.38

-1.35

-1.34

-1.33

POSITIVE LOGITS

2.42

</h2>

1.74

not

1.64

</h4>

1.61

 about

1.56

 something

1.56

YECTO

1.53

</b>

1.52

yship

1.52

layed

1.51

Activations Density 0.004%