INDEX

Explanations

everyone ! circle back majority source

specific concepts and objects

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

ва

0.54

天文

0.47

Boulder

0.47

лия

0.46

 кван

0.46

за

0.45

ра

0.45

Sandra

0.45

侔

0.44

Homme

0.43

POSITIVE LOGITS

0.57

0.49

gef

0.49

0.48

 front

0.47

 Pern

0.47

 ailes

0.47

 enemigos

0.47

 aplicaciones

0.47

Tip

0.47

Activations Density 0.000%