INDEX

Explanations

joking

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

 Laugh

-0.15

 mockery

-0.14

 mocked

-0.12

 amused

-0.11

 mocking

-0.10

 Junk

-0.10

æĥĳ

-0.10

goog

-0.09

 laughing

-0.09

 ridicule

-0.09

POSITIVE LOGITS

 kidding

0.29

 joke

0.28

pun

0.23

rib

0.23

 jokes

0.23

 jest

0.23

 facet

0.19

 cracking

0.18

wis

0.18

pun

0.18

Activations Density 0.225%