INDEX

Explanations

permutations and combinations

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

enga

-0.11

Bil

-0.11

AME

-0.10

bil

-0.10

 concaten

-0.09

 rational

-0.09

 pairwise

-0.09

 Furious

-0.09

ebra

-0.09

imus

-0.09

POSITIVE LOGITS

 combinations

0.26

binations

0.21

 comb

0.21

 Combination

0.19

 combination

0.19

_comb

0.17

 perm

0.16

Comb

0.16

ç¬

0.16

combination

0.16

Activations Density 0.091%