INDEX

Explanations

created or built

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

-held

-0.11

 devis

-0.11

-elected

-0.09

Introduced

-0.09

/Instruction

-0.09

-used

-0.09

kyt

-0.09

 CrossAxisAlignment

-0.09

huy

-0.08

.jackson

-0.08

POSITIVE LOGITS

 created

0.32

 produced

0.25

created

0.24

 creation

0.22

 Created

0.21

Created

0.21

 cread

0.19

 developed

0.19

 ÑģÐ¾Ð·Ð´

0.19

 built

0.18

Activations Density 0.224%