INDEX

Explanations

periods, phases, chapters

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

 timelines

-0.09

iren

-0.09

viar

-0.08

(workspace

-0.08

olini

-0.08

 trick

-0.08

ONO

-0.08

IDX

-0.08

trag

-0.08

POSITIVE LOGITS

 period

0.31

 phase

0.26

 chapter

0.25

 stage

0.22

 experience

0.21

 journey

0.19

 chapters

0.19

éĺ¶æ®µ

0.18

phase

0.18

period

0.18

Activations Density 0.121%