INDEX

Explanations

rewards and incentives

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gang

-0.10

åĲ¸åıĸ

-0.10

 tÃł

-0.10

æ±²åıĸ

-0.09

æĬĬæı¡

-0.09

ä¾£

-0.09

Ð¿Ð¸ÑĤ

-0.09

ross

-0.09

Specifier

-0.09

æĬķèµĦé¢Ŀ

-0.08

POSITIVE LOGITS

å¥ĸåĬ±

0.29

 rewards

0.22

 reward

0.22

å¥ĸ

0.18

 Rewards

0.18

çįİ

0.18

çĬĴ

0.18

Reward

0.17

 Reward

0.17

reward

0.17

Activations Density 0.066%