INDEX

Explanations

cracks, chips, and tears

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

 verhind

-1.34

 müm

-1.30

 hemel

-1.21

 INFORMACIÓN

-1.20

 sejam

-1.18

 TÉCN

-1.17

 gebre

-1.17

עה

-1.17

 walang

-1.16

 reactie

-1.16

POSITIVE LOGITS

部分

1.22

 раздел

1.16

 недавно

1.15

】：

1.14

 poziomie

1.11

 части

1.09

 Several

1.08

已经

1.05

𝑈

1.05

 Substanz

1.04

Activations Density 0.054%