INDEX

Explanations

commas followed by conjunctions

signaling contrast

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

،

0.39

，

0.35

0.32

ও

0.28

、

0.28

 optimisation

0.27

及

0.26

τος

0.26

而言

0.25

 във

0.25

POSITIVE LOGITS

 prompting

0.55

 meaning

0.52

 necessitating

0.52

allowing

0.51

 whereupon

0.50

 resulting

0.48

 allowing

0.48

resulting

0.47

meaning

0.46

 emphasizing

0.45

Activations Density 0.990%