INDEX

Explanations

withstand or withhold

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

داية

-2.28

𓂅

-2.27

 termed

-2.25

Going

-2.20

 grievance

-2.16

 follows

-2.11

Getting

-2.09

because

-2.08

 стал

-2.08

 devia

-2.06

POSITIVE LOGITS

霂

2.52

ers

2.47

を使用した

2.39

。“

2.36

ᕱ

2.23

так

2.22

寕

2.17

ce

2.16

潆

2.11

偁

2.09

Activations Density 0.003%