INDEX

Explanations

also provides funding or has

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

 supervise

-0.09

 allows

-0.09

 RELATED

-0.09

erin

-0.08

Dah

-0.08

ACKET

-0.08

omics

-0.08

ihan

-0.08

allows

-0.08

/internal

-0.08

POSITIVE LOGITS

 also

0.14

act

0.13

 working

0.11

 work

0.11

 cung

0.11

 tail

0.11

 provide

0.10

also

0.10

 publishing

0.10

 played

0.10

Activations Density 0.084%