INDEX

Explanations

integrated frameworks and services

This neuron detects capitalized key terms—typically the prominent adjectives or nouns used in formal names, titles, or headings (e.g. “Open,” “Unified,” “Joint,” “Consolidated,” “Comprehensive,” “Integrated,” etc.).

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

隳

-1.95

 aranha

-1.92

⮑

-1.84

 goût

-1.78

 zoude

-1.77

 aguja

-1.73

 cerâmica

-1.71

ſſen

-1.69

 delà

-1.68

僦

-1.68

POSITIVE LOGITS

 Very

1.85

There

1.81

in

1.74

 What

1.66

 even

1.62

just

1.62

If

1.62

for

1.61

or

1.60

 Even

1.59

Activations Density 0.208%