INDEX

Explanations

comparisons

comparison using "than"

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

ERS

0.55

 waahanga

0.48

 acontecer

0.46

 minhas

0.45

 raggiunto

0.45

 attraverso

0.44

 necessidades

0.44

iciado

0.44

 disparu

0.44

 chamado

0.43

POSITIVE LOGITS

than

0.60

 Than

0.59

。

0.54

 than

0.52

し

0.50

Than

0.49

än

0.48

前者

0.47

ü

0.45

 než

0.44

Activations Density 0.030%