INDEX

Explanations

construction industry work

This neuron activates on capitalized section or list item titles and labels (e.g. headings and numbered/legal references).

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

pantal

-1.77

pantalon

-1.75

figurine

-1.70

playmobil

-1.63

そこから

-1.61

θρώ

-1.59

 černá

-1.57

唖

-1.57

 cinturón

-1.57

༛

-1.55

POSITIVE LOGITS

 насеља

1.81

虽然

1.75

ׇ

1.66

如果

1.61

 државе

1.59

Если

1.59

 with

1.56

除了

1.54

雖然

1.45

「

1.42

Activations Density 0.003%