INDEX

Explanations

properly configured or added

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

 Laurent

-0.10

-0.09

 âŁ

-0.09

 compreh

-0.09

anal

-0.09

 Berk

-0.08

Rex

-0.08

 Sylv

-0.08

addon

-0.08

zen

-0.08

POSITIVE LOGITS

 correct

0.24

 spell

0.24

 spelling

0.23

æŃ£ç¡®

0.22

 spelled

0.21

 correctly

0.19

spell

0.19

 Spell

0.19

 Ð¿ÑĢÐ°Ð²Ð¸Ð»ÑĮÐ½Ð¾

0.18

 spells

0.18

Activations Density 0.045%