INDEX

Explanations

using System

The neuron fires on import‐style namespace references to “System,” i.e. the “using System…” lines.

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

 hjäl

-1.80

 heighten

-1.75

 being

-1.66

ÃO

-1.61

癇

-1.58

 alebo

-1.56

 adaptar

-1.54

lgari

-1.53

 Sahne

-1.53

 energetic

-1.52

POSITIVE LOGITS

of

2.61

The

2.34

↵

2.13

↵↵

2.11

</i>

2.02

In

1.94

.,

1.75

More

1.70

Or

1.61

<td>

1.60

Activations Density 0.007%