INDEX

Explanations

within the

mechanisms for learning

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

 gorillas

0.50

-=

0.45

 wasps

0.44

 rescuers

0.43

 speed

0.42

 mammals

0.42

ᅧ

0.42

 chests

0.42

的速度

0.42

ische

0.42

POSITIVE LOGITS

juč

0.51

ồng

0.50

 Políticas

0.50

 Fundação

0.49

ні

0.49

 Satisf

0.49

ेन्द्र

0.48

 Abschluss

0.48

 ولم

0.47

 Elemental

0.46

Activations Density 0.002%