INDEX

Explanations

originating from within

The neuron is detecting instances of the word “within.”

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

as

-2.02

 energética

-1.88

</em>

-1.88

-1.77

</td>

-1.76

 morgon

-1.71

 veckan

-1.62

了几

-1.60

 eléctrica

-1.59

 hidrá

-1.59

POSITIVE LOGITS

of

2.61

 also

2.09

1.91

 mereka

1.87

नलोड

1.85

沨

1.75

 They

1.73

вання

1.70

 также

1.63

 اعضای

1.61

Activations Density 0.010%