INDEX

Explanations

eating and drinking contexts

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

 bakery

-0.11

 Starbucks

-0.11

 pastry

-0.11

 chocolate

-0.11

 candy

-0.10

 Chocolate

-0.10

 McDonald

-0.10

 chocol

-0.10

 cash

-0.09

 baker

-0.09

POSITIVE LOGITS

 table

0.15

 eating

0.15

 enjoying

0.15

 enjoy

0.14

eat

0.14

 Dinner

0.13

 seated

0.13

 eats

0.13

ate

0.13

 enjoyed

0.13

Activations Density 0.058%