INDEX

Explanations

registering asylum seekers

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

 captive

-0.11

opr

-0.10

 prisoner

-0.10

 jails

-0.10

ç®

-0.10

æĨ

-0.09

ä¾į

-0.09

abble

-0.09

Lan

-0.09

rego

-0.09

POSITIVE LOGITS

 asylum

0.24

ylum

0.19

ASY

0.19

 processing

0.17

Ref

0.17

ref

0.17

-ref

0.16

 refuge

0.16

 Dublin

0.16

 reception

0.16

Activations Density 0.020%