INDEX

Explanations

REST API, transaction management

The neuron activates on proper‐noun tokens—particularly acronyms and named entities (e.g. technical component names, library or project acronyms, and place or organization names).

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

솜

0.64

 considerate

0.52

ffle

0.51

NCH

0.51

蒽

0.51

 пар

0.50

স্ব

0.50

ర్ణ

0.50

 unsettling

0.50

 forbind

0.50

POSITIVE LOGITS

 javax

0.86

 Caribbean

0.83

="#{

0.78

Caribbean

0.78

 Barbados

0.76

沖縄

0.75

javax

0.74

ejb

0.73

 沖縄

0.72

 Faces

0.72

Activations Density 0.058%