INDEX

Explanations

engineering assistant or design

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

arah

-0.10

------------

-0.10

 deferred

-0.09

ãģŀ

-0.09

 Bast

-0.09

alytics

-0.09

oure

-0.08

carousel

-0.08

ErrorException

-0.08

pga

-0.08

POSITIVE LOGITS

ered

0.21

å¸Ī

0.19

å¸«

0.17

ERING

0.17

ings

0.14

ering

0.14

 feats

0.13

erd

0.12

 marvel

0.12

ãĥªãĥ³ãĤ°

0.12

Activations Density 0.023%