INDEX

Explanations

work and results

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 մաս

0.59

텟

0.58

寸

0.57

 tapes

0.57

➤

0.57

erland

0.56

abeth

0.54

 राजमार्ग

0.54

 নিলাম

0.54

‣

0.54

POSITIVE LOGITS

 valida

0.54

Fabric

0.54

War

0.53

report

0.53

velocity

0.53

iki

0.52

 válida

0.52

Tit

0.51

Report

0.51

мани

0.51

Activations Density 0.000%

This feature has no known activations.