INDEX

Explanations

tim and words starting with tim

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

esan

-0.11

erence

-0.10

ITA

-0.10

/filepath

-0.10

wr

-0.10

velle

-0.10

ered

-0.09

TimeString

-0.09

ãĢģãĢģ

-0.09

snd

-0.09

POSITIVE LOGITS

othy

0.20

tim

0.19

tim

0.18

idity

0.17

Tim

0.17

bers

0.17

my

0.15

Tim

0.14

eliness

0.14

ely

0.14

Activations Density 0.019%