INDEX

Explanations

supporting or advocating for

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 disagreements

-0.11

 scandals

-0.10

Jur

-0.10

 disagree

-0.10

 Holiday

-0.10

 Crosby

-0.10

aland

-0.10

 liberated

-0.10

 threat

-0.09

 shortage

-0.09

POSITIVE LOGITS

 against

0.13

 keeping

0.13

/op

0.13

 retaining

0.12

 retention

0.12

 inclusion

0.12

against

0.12

 equality

0.12

 continued

0.12

 rights

0.12

Activations Density 0.394%