INDEX

Explanations

functional or operational states

explaining why something works

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

 LEVEL

0.38

of

0.37

 charge

0.35

of

0.34

 legger

0.34

rating

0.34

HX

0.33

 சார்பில்

0.33

POSITIVE LOGITS

 pleinement

0.49

astă

0.46

uelos

0.44

 fortemente

0.41

നില്‍

0.40

 dearly

0.39

তনে

0.38

 ως

0.37

 montón

0.37

кара

0.37

Activations Density 0.172%