INDEX

Explanations

taking pleasure

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

 å¿

-0.11

ipt

-0.10

 deepcopy

-0.10

ingham

-0.09

emean

-0.09

Zam

-0.09

IMIT

-0.09

ursal

-0.09

 homage

-0.09

ãģŁãģı

-0.08

POSITIVE LOGITS

 pleasure

0.42

 enjoyment

0.34

 satisfaction

0.32

 enjoy

0.31

 delight

0.30

äº«

0.29

 plaisir

0.28

joy

0.27

 Enjoy

0.27

 enjoys

0.27

Activations Density 0.116%