INDEX

Explanations

specific problem or task

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

erie

-0.09

Worksheet

-0.08

 Ulus

-0.08

_TC

-0.08

èµĦæł¼

-0.08

 Ð±ÑĥÐ¼Ð°

-0.08

agara

-0.08

POSITIVE LOGITS

 problem

0.39

 task

0.37

 tasks

0.30

problem

0.28

 Problem

0.28

 problems

0.26

ä»»åĬ¡

0.26

Problem

0.25

éĹ®é¢ĺ

0.25

 problema

0.25

Activations Density 0.181%