INDEX

Explanations

class definition keywords

The neuron detects code identifiers—particularly variable or field names (e.g. “file,” “key,” “date”) in declarations.

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

 miliki

-1.70

we

-1.63

 blanches

-1.62

ꪹ

-1.59

and

-1.58

และ

-1.57

relatively

-1.57

primarily

-1.55

 чрезвы

-1.53

órica

-1.51

POSITIVE LOGITS

</h2>

1.93

<bos>

1.80

 для

1.61

If

1.57

There

1.54

You

1.49

hülle

1.46

↵↵

1.43

га

1.41

Activations Density 0.085%