INDEX

Explanations

comparing relative change next year

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

:.:

-0.10

oui

-0.09

Ley

-0.09

ï¾ī

-0.09

 Patri

-0.09

 raised

-0.09

å±Ģ

-0.08

kova

-0.08

onis

-0.08

XXXX

-0.08

POSITIVE LOGITS

 compared

0.30

 Compared

0.20

comp

0.16

 year

0.15

 relative

0.15

RelativeTo

0.13

 Ð¿Ð¾ÑĢÑĸÐ²

0.13

(comp

0.12

pared

0.12

 overall

0.12

Activations Density 0.052%