INDEX

Explanations

environment and actions

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

ãĥ¾

-0.11

Yen

-0.10

 Jasper

-0.09

ieu

-0.09

IFA

-0.08

å¿µ

-0.08

èĬ¯

-0.08

zb

-0.08

AMED

-0.08

neys

-0.08

POSITIVE LOGITS

gym

0.25

Gym

0.24

 agents

0.19

env

0.18

 agent

0.18

 environment

0.17

 Agents

0.17

 environments

0.16

 Agent

0.16

Muj

0.16

Activations Density 0.043%