INDEX

Explanations

dealing with relevance

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ieri

-0.09

pollo

-0.08

 Flour

-0.08

inka

-0.08

 domic

-0.07

 sincer

-0.07

icker

-0.07

aver

-0.07

 privately

-0.07

POSITIVE LOGITS

 relevant

0.74

 Relevant

0.65

relevant

0.63

 relevance

0.62

 related

0.56

 pertinent

0.56

 relev

0.55

 relate

0.45

related

0.45

 relating

0.44

Activations Density 2.093%