INDEX

Explanations

built-in

The neuron fires strongly on instances of the phrase “built-in” (including hyphenated or split forms), i.e. mentions of something being built in.

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

ごめん

-1.86

</h3>

-1.84

美国的

-1.64

 utiliz

-1.61

老太太

-1.55

excellent

-1.53

伟大

-1.51

鱚

-1.51

 citing

-1.48

POSITIVE LOGITS

〟

1.86

,"

1.83

)$,

1.80

谞

1.74

oluted

1.71

nipeg

1.70

就

1.70

 证书

1.69

and

1.67

此外

1.67

Activations Density 0.007%