INDEX

Explanations

end of a sequence or period

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

åºı

-0.10

è½¬è½½

-0.09

ä¸įæŃ¢

-0.09

è°©

-0.09

olk

-0.08

å¢ģ

-0.08

ä¸ĭåįĬ

-0.08

 Result

-0.08

ä¸įä½Ĩ

-0.08

é¦ĸä½į

-0.08

POSITIVE LOGITS

æľĢåĲİä¸Ģæ¬¡

0.13

(before

0.10

before

0.10

 antes

0.09

Before

0.09

Touches

0.09

 before

0.09

çľĭä¸Ģçľĭ

0.08

	before

0.08

_before

0.08

Activations Density 0.042%