INDEX

Explanations

japan and japanese culture

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

ëĶĶìĭľ

-0.11

Â©Â©

-0.11

 typealias

-0.10

toi

-0.10

 aalborg

-0.10

ãĨ

-0.10

@nate

-0.10

_exempt

-0.10

<|begin_of_text|>

-0.10

oshi

-0.10

POSITIVE LOGITS

 Báº£n

0.20

eses

0.16

 Alps

0.13

ase

0.13

ise

0.12

ophile

0.12

ese

0.11

imation

0.11

 Diet

0.11

 culture

0.11

Activations Density 0.040%