INDEX

Explanations

calibration process

The neuron is triggered by occurrences of the technical term “calibration” (and close morphological variants) in measurement or setup contexts.

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

seamnă

-2.16

</h3>

-2.00

Hva

-1.95

annica

-1.87

</h2>

-1.85

 všech

-1.83

᠂

-1.83



-1.82

 quintessential

-1.82

 pueden

-1.82

POSITIVE LOGITS

the

2.16

谞

1.95

ケース

1.91

.’”

1.91

From

1.85

 festa

1.85

ではまた

1.79

For

1.78

—,

1.76

of

1.74

Activations Density 0.009%