INDEX

Explanations

package names and code snippets

The neuron strongly activates on numeric tokens (digits or numbers) such as version numbers, IDs, or other standalone numbers embedded in the text.

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

-2.17

"`

-1.92

 admiten

-1.87

 první

-1.86

 принял

-1.84

 вышел

-1.79

 pasan

-1.77

 работают

-1.76

/"

-1.75

 články

-1.75

POSITIVE LOGITS

of

2.08

。“

1.94

↵↵

1.77

哢

1.77

 fria

1.71

。『

1.62

Итог

1.60

1.59

.”

1.59

This

1.58

Activations Density 0.006%