INDEX

Explanations

happy to help

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

AdminController

-0.09

âĶ´

-0.09

Tac

-0.09

ogn

-0.08

 positional

-0.08

ä¸įå¾Ĺ

-0.08

Parm

-0.08

fug

-0.08

æĢĿãģĦ

-0.08

 Gallery

-0.08

POSITIVE LOGITS

 ready

0.21

 here

0.20

 sáºµn

0.19

 pleasure

0.17

 Ð³Ð¾ÑĤÐ¾Ð²

0.17

Ready

0.17

 hazÄ±r

0.17

 READY

0.17

ready

0.17

 Ready

0.16

Activations Density 0.103%