INDEX

Explanations

thoughts and thinking

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

æĬł

-0.09

vang

-0.09

çº¿ä¸Ĭ

-0.09

ember

-0.09

åİļåº¦

-0.08

_LESS

-0.08

è§Į

-0.08

 Ð²ÑģÐµÐ³Ð¾

-0.08

éħ¥

-0.08

_testing

-0.08

POSITIVE LOGITS

 thoughts

0.19

å¿µå¤´

0.18

 Thoughts

0.17

æĥ³åĪ°

0.14

thought

0.13

Thought

0.12

çļĦæĥ³æ³ķ

0.12

æĥ³æ³ķ

0.12

èĦĳåŃĲéĩĮ

0.11

æĥ³çĿĢ

0.11

Activations Density 0.065%