INDEX

Explanations

matching, execution, absorbing

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

 match

-1.33

absorption

-1.13

 execution

-1.11

 absorption

-1.10

 Absorption

-1.09

 executed

-1.06

Absorption

-1.06

GEBURTSDATUM

-1.06

 matches

-1.04

 matching

-1.03

POSITIVE LOGITS

0.57

it

0.56

gate

0.56

Wid

0.55

nit

0.54

pres

0.54

 hyper

0.54

erville

0.52

ta

0.52

wr

0.52

Activations Density 0.059%