INDEX

Explanations

setting default limits

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

Eck

-0.11

 discret

-0.09

 Murdoch

-0.09

utzt

-0.09

à¸ĩà¸¨

-0.09

Fro

-0.09

 subpoena

-0.08

 selectable

-0.08

 konkrÃ©t

-0.08

lej

-0.08

POSITIVE LOGITS

 setting

0.23

 Setting

0.19

 settings

0.17

Setting

0.17

setting

0.17

 value

0.17

è®¾ç½®

0.14

 values

0.13

åĢ¼

0.13

è¨Ńå®ļ

0.13

Activations Density 0.091%