INDEX

Explanations

conjunctions (and) in many languages

This neuron detects structured metadata fields and their values (e.g. section headers like “Category:”, “Type:”, “Action Date:”, and associated numeric codes).

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

 دونوں

0.89

 दोनों

0.73

दोनों

0.65

 zarówno

0.65

 ম্যাক

0.60

或其他

0.60

 ทั้ง

0.59

 或者

0.59

Both

0.57

 zowel

0.57

POSITIVE LOGITS

và

1.26

 ಮತ್ತು

1.18

 ਅਤੇ

1.18

и

1.17

 અને

1.16

 आणि

1.13

และ

1.12

and

1.09

 και

1.09

Và

1.06

Activations Density 0.884%