INDEX

Explanations

reflect on

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

to

-1.10

を目指す

-0.93

Popis

-0.93

낌

-0.92

rieb

-0.92

Beskrivelse

-0.90

chent

-0.88

 توضی

-0.87

 açıklama

-0.87

PARAMETER

-0.86

POSITIVE LOGITS

 upon

1.93

 about

1.58

 reflection

1.56

reflection

1.38

upon

1.38

Upon

1.35

Reflect

1.33

 Upon

1.32

 Reflection

1.31

 reflek

1.28

Activations Density 0.013%