INDEX

Explanations

visited, processed, or traversed items

The neuron spikes on code identifiers used for internal state tracking in algorithms—especially names like “stack” and “visited.”

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

 their

-1.88

to

-1.72

-1.59

-1.44

-1.42

-1.40

 current

-1.40

<strong>

-1.39

-1.34

POSITIVE LOGITS



1.69

Välisling

1.55

 ヌード

1.50

Menge

1.48

 варто

1.45

 moft

1.41

ሣ

1.39

 Pokud

1.38

Aten

1.38

 neceffary

1.37

Activations Density 0.007%