INDEX

Explanations

attached documents or letters

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 spoken

-0.14

 verbal

-0.13

 speeches

-0.13

 audible

-0.13

 whispered

-0.13

 conversation

-0.12

 conversations

-0.12

 vocal

-0.11

 Conversation

-0.11

 whispers

-0.10

POSITIVE LOGITS

 wrote

0.20

 enclosed

0.19

 writing

0.18

 included

0.18

 therein

0.18

 Included

0.17

 paragraph

0.17

 paragraphs

0.17

 Attached

0.16

 writes

0.16

Activations Density 0.275%

No Known Activations