INDEX

Explanations

song titles

The neuron strongly activates on isolated, capitalized single words that serve as section or title headings.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

The

0.41

Festival

0.38

THE

0.37

FREE

0.36

Ра

0.35

О

0.34

 masterpiece

0.33

mystery

0.33

В

0.33

Halloween

0.33

POSITIVE LOGITS

 كويس

0.37

 Nascimento

0.36

 hypothalamic

0.35

 namespaces

0.35

 olduğuna

0.34

 माइक्रोस

0.33

 dışında

0.33

 Sedangkan

0.33

 वेक्टर

0.32

 valign

0.32

Activations Density 0.001%