INDEX

Explanations

completing text or tasks

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

ye

-0.10

 Brook

-0.10

 entr

-0.09

gag

-0.08

portun

-0.08

ajs

-0.08

Mun

-0.08

ienza

-0.08

opy

-0.08

 Carm

-0.08

POSITIVE LOGITS

 completion

0.12

 Completion

0.12

å®ĮæĪĲ

0.10

 Comple

0.10

udden

0.10

Completion

0.10

 completing

0.10

_completion

0.09

 flesh

0.09

ÐĿÐĨ

0.09

Activations Density 0.045%