INDEX

Explanations

the abstract concept

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

highest

2.80

 heaviest

2.79

shortest

2.70

 shortest

2.69

exact

2.65

 maximale

2.64

 coolest

2.61

 maksymal

2.56

 southernmost

2.52

 tallest

2.51

POSITIVE LOGITS

 affaire

1.67

 demain

1.66

 questione

1.39

 aujourd

1.39

 reconsider

1.38

 rethink

1.33

 dilemma

1.28

Proble

1.27

Agenda

1.25

 crisis

1.23

Activations Density 0.173%

No Known Activations