INDEX

Explanations

ML parameters and lists

This neuron responds to capitalized, title‐case tokens such as proper nouns, headings, or list‐item labels.

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

 Mandarin

0.86

 Sometimes

0.83

 Illustrations

0.83

 ANALYSIS

0.82

 excerpts

0.82

 Algunas

0.81

 Painting

0.81

 Mél

0.79

 ILLUSTR

0.79

 ASSESSMENT

0.79

POSITIVE LOGITS

0.74

 admiss

0.74

ায়

0.71

خاب

0.70

 propriedade

0.70

ingress

0.70

зия

0.69

斂

0.69

 பயன்ப

0.67

тности

0.67

Activations Density 0.001%