INDEX

Explanations

requested parameters and temperature

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

Sne

-0.09

actice

-0.08

ograd

-0.08

 family

-0.08

olf

-0.08

ice

-0.08

ersh

-0.08

 ÙĦØº

-0.08

 Stir

-0.08

fam

-0.08

POSITIVE LOGITS

 conditions

0.21

conditions

0.17

 requirements

0.16

 environmental

0.16

 clim

0.16

 Conditions

0.15

/weather

0.15

 answers

0.14

 weather

0.14

 climate

0.14

Activations Density 0.051%

No Known Activations