INDEX

Explanations

can show

reveals or tells

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 Promoting

0.43

 പ്രവർ

0.43

 boosting

0.42

 causando

0.41

promoting

0.41

 promover

0.40

 promoting

0.40

causing

0.40

 تحقق

0.39

Accom

0.39

POSITIVE LOGITS

 revealing

1.84

 reveal

1.76

 reveals

1.76

 Reveals

1.61

reveal

1.59

 Reveal

1.51

reve

1.48

 revelar

1.44

Reve

1.39

Reveal

1.39

Activations Density 0.148%