INDEX

Explanations

numerical decrease

values are low or below

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 yüksek

1.06

 høj

1.06

 høy

1.03

 उच्च

0.99

 высокий

0.98

 υψη

0.97

 higher

0.96

ขึ้น

0.96

 높은

0.96

 höher

0.96

POSITIVE LOGITS

 below

1.31

 ниже

1.29

 아래

1.20

below

1.20

 नीचे

1.15

Below

1.10

 aşağı

1.10

 BELOW

1.10

 নিচে

1.09

 Below

1.05

Activations Density 0.064%