INDEX

Explanations

sequence continuation or outcome

assessment and considerations

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 plains

0.44

व्हाण

0.44

வ்வாறு

0.43

鎿

0.43

अगर

0.42

嬰

0.42

솝

0.42

銨

0.42

लेकिन

0.42

अपनी

0.41

POSITIVE LOGITS

 chce

0.55

 danach

0.53

 fehlt

0.53

״

0.52

tte

0.51

 idem

0.51

}{

0.50

 leichte

0.50

いたします

0.49

 einzig

0.49

Activations Density 0.002%