INDEX

Explanations

learned behavior, muscle, strengthening

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

Tac

-0.10

é¢Ĩ

-0.09

èĦ±

-0.09

åħĪéĶĭ

-0.09

åĪ¨

-0.08

æį£

-0.08

Tac

-0.08

èµĦæºĲä¼ĺåĬ¿

-0.08

åīª

-0.08

éĩľ

-0.08

POSITIVE LOGITS

éĶ»çĤ¼

0.14

ç»ĥä¹ł

0.14

ç»ĥ

0.12

ä¼¸

0.12

çŃĭ

0.11

ç·´

0.11

å¼¥

0.10

 exercised

0.10

éĶ»

0.10

åľ°çĲĥä¸Ĭ

0.10

Activations Density 0.059%