INDEX

Explanations

declining harmful requests

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

alu

-0.10

å»Ĭ

-0.09

ucc

-0.09

oku

-0.08

 Blitz

-0.08

 nods

-0.08

 grou

-0.08

 exem

-0.08

antal

-0.08

oru

-0.08

POSITIVE LOGITS

 sorry

0.12

 unable

0.12

 cannot

0.11

 Sorry

0.10

 declining

0.10

cannot

0.10

ANNOT

0.10

 inability

0.10

sorry

0.10

 refused

0.09

Activations Density 0.052%