INDEX

Explanations

Service class instantiation

The neuron fires on occurrences of the identifier “service” (especially the capitalized token “Service”) in code—e.g. imports, annotations, class or variable names.

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

‘

-2.11

而

-1.86

</h3>

-1.77

});

-1.77

}$,

-1.76

?”

-1.73

…

-1.70

):

-1.70

 weshalb

-1.68

 '-':

-1.67

POSITIVE LOGITS

 parachoque

2.19

そういえば

2.08

枱

2.05

明日は

1.98



1.93

當初

1.92

それにしても

1.89

素晴らしい

1.87

というわけで

1.85

囝

1.85

Activations Density 0.005%