INDEX

Explanations

cause/reason conjunctions

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

å¯¹ä½ł

-0.06

isia

-0.06

Ã¤nde

-0.06

ä½ĳ

-0.06

reto

-0.06

 long

-0.06

reur

-0.06

è¿Ļæĺ¯æĪĳçļĦ

-0.06

åįĥ

-0.06

amer

-0.06

POSITIVE LOGITS

çĶ±äºİ

0.15

 Since

0.15

Since

0.15

çĶ±æĸ¼

0.15

since

0.14

 since

0.14

çĶ±äºİåħ¶

0.14

æĹ¢çĦ¶

0.14

 ÐŁÐ¾ÑģÐºÐ¾Ð»ÑĮÐºÑĥ

0.12

åĽłä¸º

0.11

Activations Density 0.027%