INDEX

Explanations

discussion or decisions

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

ubar

-0.10

 writ

-0.09

 Interviews

-0.09

 SOUR

-0.09

 interviews

-0.08

_reply

-0.08

à¤ľà¤°

-0.08

Reply

-0.08

 essay

-0.08

 Essay

-0.08

POSITIVE LOGITS

 discussion

0.29

 discussed

0.29

 discuss

0.27

 Ð¾Ð±Ñģ

0.26

è®¨

0.25

 agenda

0.24

 discussing

0.23

 discusses

0.22

è¨İ

0.22

 Discussion

0.21

Activations Density 0.131%