INDEX

Explanations

resembling shapes and appearances

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

 Robbins

-0.09

 Harr

-0.09

ets

-0.09

 insets

-0.09

 mell

-0.09

idget

-0.09

NOP

-0.09

 Mell

-0.09

adam

-0.08

belongs

-0.08

POSITIVE LOGITS

 resemble

0.38

 resembling

0.38

 resembles

0.37

 resembled

0.36

 resemblance

0.33

 resembl

0.29

 looks

0.22

 liken

0.22

 likeness

0.20

looks

0.20

Activations Density 0.164%