INDEX

Explanations

Here's why this is problematic

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

 mempunyai

2.90

 보면은

2.31

非常的

2.15

 mentioned

2.14

這樣子

2.13

 pursuance

2.09

有一定的

2.06

ぐらい

2.05

種類

2.04

 equipments

2.03

POSITIVE LOGITS

即便

1.73

ớt

1.72

ực

1.69

晚

1.68

 nuanced

1.61

 प्रतिबद्ध

1.59

런치

1.59

ốt

1.59

 cuối

1.58

熠

1.57

Activations Density 0.314%