INDEX

Explanations

she said, actually a, alone lift

potentially lose

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

莱

0.56

 अद

0.50

 व्यापारियों

0.49

 राजकुमारी

0.47

ໂດ

0.46

ogenen

0.46

ով

0.46

स्ताना

0.46

 raiding

0.46

 lẻ

0.45

POSITIVE LOGITS

 bowl

0.42

するもの

0.40

Goa

0.40

するなど

0.40

hap

0.39

Universidad

0.38

ষা

0.38

Hap

0.38

 propensity

0.37

Activations Density 0.005%