INDEX

Explanations

units and measurements

The neuron strongly activates on standalone numeric tokens (e.g. literal digits or numbers in code).

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

 euró

-1.12

 vestidos

-1.05

striker

-1.00

stags

-0.98

Greek

-0.97

almo

-0.95

armani

-0.95

 ekolog

-0.95

ventus

-0.94

roach

-0.93

POSITIVE LOGITS

can

1.05

TECN

0.96

val

0.95

Simpli

0.94

0.93

mon

0.92

 Поэтому

0.90

ze

0.89

will

0.89

Activations Density 0.002%