INDEX

Explanations

consistent with and inconsistent

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

 devenu

-1.03

 compleja

-0.97

 Misalnya

-0.96

 sekitar

-0.93

 aktiv

-0.90

 chargement

-0.88

elm

-0.87

 kegiatan

-0.87

andai

-0.86

ógy

-0.86

POSITIVE LOGITS

 Consistency

1.80

consistency

1.71

consistent

1.64

 consistency

1.54

 inconsistency

1.48

Consistency

1.45

 Consistent

1.43

 consistent

1.43

 inconsistencies

1.42

 inconsistent

1.41

Activations Density 0.028%