INDEX

Explanations

A&M, A&E, A&P

The neuron detects the ampersand character (“&”) when used to join abbreviated names or entities (e.g. “A & M,” “A & E”).

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

 feitos

-2.33

is

-2.11

’

-2.06

-1.86

-1.77

_{

-1.77

結び

-1.76

 позволит

-1.76

᠂

-1.72

𐑂

-1.67

POSITIVE LOGITS

枓

1.88

 löytyy

1.73

？」

1.72

 esser

1.70

 Inſ

1.69

⸎

1.67

不是

1.67

大部分

1.66

 naturali

1.64

 purview

1.63

Activations Density 0.003%