INDEX

Explanations

impact and influence

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

çº¸ä¸Ĭ

-0.10

hab

-0.09

ä¼ĺ

-0.09

æĬ¢

-0.09

ìķ¼

-0.09

åħ¨ç¨ĭ

-0.09

 ìĪĺì§ĳ

-0.08

éĩĩéĽĨ

-0.08

 vign

-0.08

è¶ĬæĿ¥è¶Ĭ

-0.08

POSITIVE LOGITS

å½±åĵį

0.15

å½±åĵįåĬĽ

0.15

çļĦå½±åĵį

0.15

å½±éŁ¿

0.14

 influence

0.13

å½±åĵįåĬĽçļĦ

0.12

 Influence

0.11

 Ð²Ð»Ð¸Ñı

0.10

 impact

0.10

æ·±è¿ľ

0.10

Activations Density 0.052%