INDEX

Explanations

references to the word "the," which is primarily used to identify and focus on specific objects, concepts, or entities within a text

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

The

-0.88

-0.77

so

-0.75

-0.73

in

-0.70

che

-0.69

 part

-0.68

sub

-0.68

…

-0.68

POSITIVE LOGITS

 Majefty

1.38

 raiſ

1.30

 Monfieur

1.29

 ſever

1.25

 myſelf

1.25

 purpoſe

1.24

 themſelves

1.23

 poffible

1.20

 avoient

1.19

Efq

1.19

Activations Density 0.367%