INDEX

Explanations

lack of necessity

redundant or unnecessary concepts

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

 lệch

0.42

मॉडल

0.40

 ansatz

0.39

্কে

0.38

最初に

0.38

 interrupted

0.38

ອ

0.37

幇

0.37

탁

0.37

 неправи

0.36

POSITIVE LOGITS

 unnecessary

1.86

 superfluous

1.52

不需要

1.45

 unnecessarily

1.37

 अनावश्यक

1.30

 needlessly

1.30

无需

1.27

 inutile

1.25

無需

1.23

不必

1.19

Activations Density 0.644%