INDEX

Explanations

code creates

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

çŁ¢

-0.10

Wr

-0.09

wr

-0.09

efa

-0.08

reb

-0.08

jas

-0.08

 Opcode

-0.08

áº¤

-0.08

kovi

-0.08

 considering

-0.08

POSITIVE LOGITS

 creates

0.23

 create

0.22

 creating

0.21

 táº¡o

0.20

 Creates

0.20

åĪĽå»º

0.20

 Create

0.18

 Creating

0.18

 åĪĽå»º

0.18

.create

0.17

Activations Density 0.136%