INDEX

Explanations

design pattern

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

 trial

-0.86

 trials

-0.77

ีย

-0.76

 campanha

-0.75

Inclusive

-0.72

 repaid

-0.71

 instinctively

-0.71

の状態

-0.68

 analogy

-0.68

Exponent

-0.68

POSITIVE LOGITS

 Variance

1.34

Variance

1.16

variance

1.16

 variances

1.15

 variance

1.14

Budget

1.03

budget

0.97

 budget

0.92

 Budget

0.90

Spi

0.88

Activations Density 0.006%