INDEX

Explanations

's important, essential, worth

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

å¤§

-0.11

å°ıè¯´

-0.11

å¤§éĩı

-0.11

äºĭ

-0.10

’;t

-0.10

è¯Ŀ

-0.10

sey

-0.09

å£°éŁ³

-0.09

äººæīį

-0.09

æĹ¶åĢĻ

-0.09

POSITIVE LOGITS

 been

0.14

ÂĢÂ

0.14

 important

0.14

 worth

0.13

ÂĿ

0.12

ÂĢÂĻ

0.12

outh

0.12

important

0.11

 importante

0.11

 Been

0.11

Activations Density 0.056%