INDEX

Explanations

provide feedback

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

 Recomm

-0.10

mÃ©na

-0.09

enet

-0.09

enza

-0.09

Fet

-0.09

BindingUtil

-0.09

_OVERRIDE

-0.08

olest

-0.08

è¡

-0.08

EMPLARY

-0.08

POSITIVE LOGITS

 feedback

0.88

 Feedback

0.77

feedback

0.71

Feedback

0.70

-feedback

0.58

_feedback

0.57

.feedback

0.52

åĽŀ

0.29

 comments

0.28

 feed

0.28

Activations Density 0.123%