INDEX

Explanations

proper nouns/abbreviations

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

MED

-0.80

hardt

-0.68

MED

-0.60

noaa

-0.58

idine

-0.56

 Bede

-0.56

sing

-0.54

titleMargin

-0.52

mit

-0.52

song

-0.52

POSITIVE LOGITS

 تانيه

0.67

 themſelves

0.66

 Infórmanos

0.62

ROK

0.59

EventLoop

0.58

 himſelf

0.57

 utafitiHapana

0.57

tonsoft

0.56

juice

0.54

Eſ

0.54

Activations Density 0.946%