INDEX

Explanations

descriptive adjectives modifying nouns

The neuron fires strongly on the initial, capitalized topic or tag word (e.g. “Exponential,” “Elastic,” “Vertical,” “Inertial,” “Static”) that heads each question.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 sementes

-1.20

 après

-1.17

 olej

-1.17

Ça

-1.17

hô

-1.16

 castelo

-1.16

 Produtos

-1.15

 parámetros

-1.13

 ajouter

-1.12

 indiví

-1.11

POSITIVE LOGITS

or

2.14

 because

1.34

1.28

 when

1.27

 usually

1.24

 When

1.24

 There

1.12

 Unlike

1.12

 Several

1.11

 Usually

1.11

Activations Density 0.337%