INDEX

Explanations

explains what something means

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

glo

-0.09

Glo

-0.09

med

-0.09

Tp

-0.09

ï¼Įä»¥åıĬ

-0.09

 ÙħØ«ÙĦØ§

-0.09

Î¿ÏħÏĤ

-0.08

ivot

-0.08

 especially

-0.08

åĮħæĭ¬

-0.08

POSITIVE LOGITS

 meaning

0.35

meaning

0.29

 Meaning

0.24

 nghÄ©a

0.23

 means

0.23

æĦıåĳ³

0.22

-mean

0.20

means

0.19

Mean

0.18

 à¤®à¤¤à¤²à¤¬

0.17

Activations Density 0.164%