INDEX

Explanations

alarmes

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

Profiler

-0.10

spe

-0.09

 Errors

-0.09

SER

-0.09

nofollow

-0.09

 synchron

-0.08

æ¬¢

-0.08

Ser

-0.08

 Transit

-0.08

iel

-0.08

POSITIVE LOGITS

 alert

0.26

alert

0.21

 Alert

0.20

 alarm

0.19

.alert

0.17

 alerts

0.17

\talert

0.17

 warning

0.17

Alert

0.17

-alert

0.17

Activations Density 0.070%