INDEX

Explanations

listing costs and challenges

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

vac

-0.09

phan

-0.09

 scarcity

-0.08

 latter

-0.08

 increasingly

-0.08

eatures

-0.08

_triggered

-0.08

æĹıèĩªæ²»

-0.08

 cruc

-0.08

åĤĻ

-0.08

POSITIVE LOGITS

 limited

0.24

 Limited

0.20

limited

0.18

Limited

0.17

 LIMITED

0.15

æľīéĻĲ

0.13

 steep

0.13

 difficulty

0.12

imited

0.12

 dependency

0.12

Activations Density 0.046%