INDEX

Explanations

answering questions

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

áček

-0.96

ally

-0.87

onomies

-0.86

 dobry

-0.86

ura

-0.84

里

-0.84

里

-0.82

 Worte

-0.81

rootNode

-0.81

 area

-0.80

POSITIVE LOGITS

 questions

2.36

 question

1.68

questions

1.57

 kysy

1.45

 preguntas

1.44

 phones

1.38

Questions

1.38

 Fragen

1.34

QUESTIONS

1.31

 Questions

1.30

Activations Density 0.015%