INDEX

Explanations

Romanian text fragments

The neuron detects academic bibliographic reference text—formatted citations including author names/initials, journal or book titles, publication years, volume/page numbers, ISBNs/ISSNs, and similar metadata.

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

例如

0.76

्यों

0.70

angg

0.70

Gender

0.70

gender

0.69

ਵਾ

0.68

्रम

0.67

 alienated

0.67

൭

0.67

ני

0.66

POSITIVE LOGITS

 pentru

1.26

 Pentru

1.20

 unei

1.20

Pentru

1.16

 foarte

1.13

 numai

1.13

 aproximativ

1.12

și

1.11

 unui

1.10

în

1.09

Activations Density 0.015%