INDEX

Explanations

what will happen

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

偻

-0.94

Gewicht

-0.94

何か

-0.94

 PUSTAKA

-0.93

something

-0.93

evre

-0.92

either

-0.91

 diras

-0.91

 something

-0.89

itize

-0.88

POSITIVE LOGITS

 happen

2.25

 happened

2.14

 happens

1.81

 happening

1.57

 constitute

1.38

 transpired

1.23

 Happens

1.22

 constitutes

1.21

 happ

1.20

be

1.15

Activations Density 0.026%