INDEX

Explanations

deep dive into

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

ÏĨÏĮ

-0.09

ä¸įèĥľæĶ¶

-0.08

NewLabel

-0.07

aterna

-0.07

å¤±çģµ

-0.07

agliari

-0.07

 efficac

-0.07

åıĳå¸ĥåħ¬åĳĬç§°

-0.07

 Gomes

-0.07

-Encoding

-0.07

POSITIVE LOGITS

æ·±åº¦

0.12

 deeper

0.12

 depth

0.12

æ·±åº¦çļĦ

0.12

Deep

0.12

deep

0.11

 sÃ¢u

0.11

 deep

0.11

 Ð³Ð»ÑĥÐ±

0.11

æ·±åħ¥çļĦ

0.11

Activations Density 0.027%

No Known Activations