INDEX

Explanations

problems, troubles, or questions

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

 greatness

-0.11

undle

-0.09

Unt

-0.09

 verbosity

-0.09

 Verg

-0.09

Vog

-0.09

 Piet

-0.09

acz

-0.08

ORN

-0.08

 Params

-0.08

POSITIVE LOGITS

 questions

0.22

 tricky

0.22

 puzz

0.22

 problems

0.22

 dile

0.20

 perplex

0.20

 quand

0.19

 dilemma

0.19

 troubles

0.18

 queries

0.18

Activations Density 0.609%