INDEX

Explanations

thinking about how

introducing analogies

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

ipal

0.41

ಾಯ

0.40

supports

0.39

 আলম

0.39

official

0.38

ื้อ

0.38

explore

0.37

yp

0.36

უნქ

0.36

＋

0.36

POSITIVE LOGITS

 twice

1.02

 about

0.90

 fondly

0.89

 aloud

0.82

 Twice

0.81

twice

0.79

 think

0.79

 differently

0.78

 Think

0.78

 negatively

0.77

Activations Density 0.093%