INDEX

Explanations

Q: questions

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

éĦĻ

-0.21

cott

-0.20

éĶĻäºĨ

-0.20

æĳĩå¤´

-0.20

 Bundy

-0.19

Kab

-0.19

 coun

-0.19

å¨ĥ

-0.19

å·¥ä½ľæĬ¥åĳĬ

-0.18

æĮĩçĿĢ

-0.18

POSITIVE LOGITS

 Ð¼Ð¾Ð¼ÐµÐ½ÑĤ

0.21

é¢ĦçķĻ

0.20

æĢĿæĥ³æĶ¿æ²»

0.20

çĪ±ä½ł

0.20

åĪ°æĹ¶åĢĻ

0.19

ÑĥÑĪ

0.19

é¹Ĭ

0.19

 Vanderbilt

0.19

/MIT

0.19

 transitions

0.19

Activations Density 0.030%