INDEX

Explanations

thanks and advance

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 himself

-0.11

 staggering

-0.10

 Genius

-0.09

aster

-0.09

 genius

-0.09

IMO

-0.08

adients

-0.08

 apparently

-0.08

 :";↵

-0.08

osten

-0.08

POSITIVE LOGITS

 sincerely

0.17

 Thank

0.14

 kindly

0.14

Thank

0.13

 please

0.13

 thank

0.13

Please

0.12

 appreciate

0.12

 Please

0.12

incerely

0.12

Activations Density 1.830%