INDEX

Explanations

phrases like corresponding value

The neuron fires on technical terms that introduce or point to quantitative comparisons or specific data points (e.g. “corresponding,” “value,” “parent,” “analogues,” “encoding,” “intermediate”).

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

but

-1.52

 illetve

-1.49

èdent

-1.45

 valamint

-1.45

ibilité

-1.41

 Lumpur

-1.41

derung

-1.38

ibilità

-1.38

umerable

-1.38

__":

-1.38

POSITIVE LOGITS

1.70

 that

1.43

 intentó

1.39

 ruft

1.34

 what

1.25

 Что

1.23

 their

1.23

our

1.23

 delito

1.22

beho

1.21

Activations Density 0.091%