INDEX

Explanations

predictive analysis and performance

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

hitheater

-0.77

setData

-0.76

Mismatch

-0.75

 موسی

-0.73

 Alegre

-0.73

loha

-0.73

 comprom

-0.72

פרו

-0.72

greet

-0.71

 dampak

-0.71

POSITIVE LOGITS

 predictive

1.39

 scored

1.23

 scoring

1.23

 score

1.16

 Predictive

1.14

score

1.09

SAS

1.05

 scores

1.04

Scoring

1.03

 Scoring

1.00

Activations Density 0.027%