INDEX

Explanations

understanding personal experience

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

 stain

-0.08

 trails

-0.08

]↵↵↵↵

-0.08

 trail

-0.08

ein

-0.08

-St

-0.07

 pÃ©

-0.07

è¿

-0.07

ctor

-0.07

glass

-0.07

POSITIVE LOGITS

 empathy

0.16

 empath

0.16

 firsthand

0.16

 understanding

0.14

 Understanding

0.14

 understand

0.12

Understanding

0.12

vic

0.11

 outsider

0.11

 sympathetic

0.11

Activations Density 0.127%