INDEX

Explanations

descriptive 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

No Comments

Negative Logits

 Lyrics

1.43

 Humor

1.34

 poetry

1.32

 Spiritual

1.31

 Creatures

1.31

 booze

1.31

 Creature

1.31

 furry

1.29

 mensch

1.28

 Insects

1.28

POSITIVE LOGITS

 shortfall

1.72

 decommissioning

1.65

 기존

1.65

 retrofit

1.61

 residual

1.57

 incremental

1.57

 backlog

1.53

 reallocation

1.52

 recalcul

1.52

既存

1.52

Activations Density 0.997%