INDEX

Explanations

please followed by offer or query

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

lew

-0.10

mediate

-0.09

 Gale

-0.09

NamedQuery

-0.09

 Bookmark

-0.09

 addCriterion

-0.09

aly

-0.08

Reached

-0.08

-metadata

-0.08

 refrain

-0.08

POSITIVE LOGITS

 feel

0.30

feel

0.25

 Feel

0.24

Feel

0.24

 feels

0.20

 felt

0.17

 sentir

0.14

 feeling

0.13

 free

0.13

 note

0.12

Activations Density 0.038%