INDEX

Explanations

This. However

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

reon

-0.11

ä¸Ķ

-0.10

assis

-0.09

 Pais

-0.09

 optionally

-0.09

cio

-0.09

 electr

-0.08

æīĢä»¥

-0.08

.Now

-0.08

á¿

-0.08

POSITIVE LOGITS

 however

0.22

 However

0.21

However

0.19

 HOWEVER

0.19

 this

0.15

 This

0.15

ï¼Įä½Ĩæĺ¯

0.15

 Ð¾Ð´Ð½Ð°ÐºÐ¾

0.15

è¿Ļç§į

0.15

This

0.14

Activations Density 0.054%