INDEX

Explanations

mistakes and negative outcomes

document structure, case studies, and negative aspects

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

约束

0.43

Workshop

0.42

Programm

0.42

 Workshop

0.41

 Activity

0.40

 Sensor

0.40

 범위

0.38

类

0.38

Studio

0.38

或

0.38

POSITIVE LOGITS

 shitty

0.53

 disgraceful

0.53

 disgrace

0.51

 hurried

0.51

 nurses

0.50

 marzo

0.50

 miserable

0.50

 fearing

0.50

 novembre

0.50

bew

0.50

Activations Density 0.000%