INDEX

Explanations

onsen facilities, yakisoba noodles

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

ings

-0.13

less

-0.12

ly

-0.12

ally

-0.11

ss

-0.10

 lightly

-0.10

ations

-0.10

er

-0.09

ers

-0.09

POSITIVE LOGITS

-style

0.16

(sp

0.13

 style

0.12

Ìģ

0.11

/***/

0.10

astically

0.10

0.09

perator

0.09

EMPLARY

0.09

 Cumhur

0.09

Activations Density 0.206%