INDEX

Explanations

can

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

Negative Logits

canonical

-0.13

 Canonical

-0.13

Canonical

-0.12

 Cannon

-0.11

crud

-0.11

 canonical

-0.09

 Canon

-0.09

 canvas

-0.09

bcm

-0.09

canvas

-0.09

POSITIVE LOGITS

can

0.51

åı¯ä»¥

0.43

 dapat

0.38

 could

0.34

 kunnen

0.33

 kÃ¶nnen

0.32

èĥ½

0.32

à¸ªà¸²à¸¡à¸²à¸£à¸ĸ

0.32

can

0.31

èĥ½å¤Ł

0.31

Activations Density 0.310%

No Known Activations