INDEX

Explanations

suggestions for improvement or desired changes

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

Materiál

-1.10

元々

-1.05

ريب

-1.05

 chô

-1.04

URIDAD

-1.02

時点で

-1.02

 canin

-1.02

𝓙

-1.02

 geene

-1.01

tableLayoutPanel

-1.00

POSITIVE LOGITS

 more

2.28

もう少し

1.68

 better

1.65

 davantage

1.45

もっと

1.30

 some

1.28

 clearer

1.27

 بیشتری

1.24

 would

1.22

 further

1.21

Activations Density 0.020%