INDEX

Explanations

vehicle upright stability

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

 Brick

-0.11

 brick

-0.10

 shoulders

-0.09

 indirect

-0.09

 ÐºÐ¸ÑģÐ»

-0.09

 jaws

-0.09

loquent

-0.08

 Tables

-0.08

lays

-0.08

inel

-0.08

POSITIVE LOGITS

 balance

0.57

 Balance

0.48

balance

0.45

Balance

0.43

_balance

0.35

 stability

0.34

 balances

0.33

 balancing

0.33

alance

0.33

.balance

0.32

Activations Density 0.067%