INDEX

Explanations

hallmark symptom or feature

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

2.59

：

2.12

2.01

:**

1.81

:-

1.77

:<

1.73

:*

1.70

:.

1.69

.:

1.63

:[

1.59

POSITIVE LOGITS

 pathogens

1.30

 incurring

1.30

 permittee

1.23

 buildup

1.23

ფ

1.23

 pathogen

1.22

 artificially

1.20

<unused425>

1.17

inducing

1.16

<unused1729>

1.15

Activations Density 0.003%