INDEX

Explanations

comma followed by but

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

יה

1.72

ി

1.70

꽁

1.68

ه

1.64

1.63

soa

1.59

hem

1.58

ⴽ

1.58

ikhil

1.57

श्वर

1.56

POSITIVE LOGITS

্ভব

2.49

ード

2.26

 PRIMARY

2.21

েনারেল

2.21

 quin

2.20

 jobId

2.19

dSample

2.16

 HARD

2.13

ogical

2.10

猬

2.05

Activations Density 0.058%