INDEX

Explanations

sql queries

The neuron fires on bits of embedded SQL code—especially SELECT/CAS T/LOWER clauses, “AS” keywords, and quoted string literals.

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

are

-1.45

-1.43

}=\

-1.41

 arrayList

-1.30

}}

-1.30

 também

-1.29

-1.28

িনি

-1.25

↵↵↵

-1.22

|;

-1.22

POSITIVE LOGITS

 from

1.81

ниці

1.55

 passenden

1.52

有多少

1.51

 their

1.50

 when

1.50

 свої

1.43

 suitable

1.39

 during

1.37

étro

1.35

Activations Density 0.040%