INDEX

Explanations

high conductivity or graphene

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

 Lime

-0.11

 lime

-0.10

 ware

-0.10

 carn

-0.09

 oppon

-0.09

 slag

-0.09

 dong

-0.09

 Gand

-0.09

 pron

-0.08

 glasses

-0.08

POSITIVE LOGITS

 carbon

0.23

 Carbon

0.22

 graphene

0.21

 Graph

0.21

 graph

0.21

Graph

0.20

Carbon

0.18

graph

0.17

carbon

0.17

.graph

0.17

Activations Density 0.036%

No Known Activations