INDEX

Explanations

help and provide

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

dov

-0.11

iku

-0.10

neau

-0.10

 Credit

-0.10

amber

-0.10

 inde

-0.09

 gang

-0.09

 autom

-0.09

alie

-0.09

OSS

-0.09

POSITIVE LOGITS

 provide

0.10

HIR

0.10

 hope

0.10

 cung

0.10

 answer

0.10

 providing

0.10

æıĲä¾Ľ

0.10

å¸ĮæľĽ

0.09

hope

0.09

 proporcion

0.09

Activations Density 0.076%