INDEX

Explanations

design & principles

The neuron fires strongly on occurrences of the word “design” (and its close derivations).

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

1.41

ات

1.35

le

1.30

동안

1.17

 Defendants

1.15

Predicate

1.15

tà

1.13

phone

1.10

دول

1.09

غ

1.09

POSITIVE LOGITS

 设计

1.74

 diseño

1.69

 дизайна

1.66

 дизайн

1.65

 conçu

1.59

 designers

1.58

 giày

1.56

 ডিজাই

1.55

ни

1.53

 Designers

1.52

Activations Density 0.077%