INDEX

Explanations

numbers and units

The neuron is specialized to pick out numeric measurements (quantities with units or parenthesized numbers).

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

so

-1.87

and

-1.55

 even

-1.48

 still

-1.40

 genannt

-1.39

nå

-1.39

認為

-1.38

as

-1.36

 called

-1.34

 auch

-1.33

POSITIVE LOGITS

aning

1.45

aned

1.41

 Bedarf

1.39

쿰

1.38

 меня

1.38

 schlimm

1.38

 gömlek

1.37

 harten

1.36

észetes

1.33

 bonheur

1.32

Activations Density 0.051%