INDEX

Explanations

require or need to

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

.ColumnHeader

-0.09

usto

-0.09

uffy

-0.09

Balancer

-0.09

Wak

-0.08

 modular

-0.08

aya

-0.08

arella

-0.08

elop

-0.08

ADE

-0.08

POSITIVE LOGITS

 quick

0.13

quick

0.12

lot

0.12

å¤§éĩı

0.11

 quickly

0.11

 lots

0.11

 absolute

0.10

 frequent

0.10

 precision

0.10

 access

0.10

Activations Density 0.071%