INDEX

Explanations

remembering and reminders

New Auto-Interp

Top Features by Cosine Similarity

Configuration

Prompts (Dashboard)

10,000 prompts, 128 tokens each

Dataset (Dashboard)

lmsys/lmsys-chat-1m

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 homage

-0.10

 Rena

-0.09

figcaption

-0.09

ãĤģãĤĭ

-0.09

GP

-0.09

 Tribute

-0.09

anka

-0.09

Å¾en

-0.08

 sing

-0.08

 receptor

-0.08

POSITIVE LOGITS

 remember

0.68

 Remember

0.61

remember

0.61

Remember

0.59

è®°

0.55

 remembers

0.54

 remembered

0.50

 nhá»Ľ

0.46

è¨ĺ

0.44

 remembering

0.39

Activations Density 0.114%