INDEX

Explanations

suggesting possibilities or actions

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

⽅

-0.91

ShouldBe

-0.90

会被

-0.89

愷

-0.89

 auttaa

-0.89

any

-0.88

 být

-0.87

 shouldBe

-0.87

'&:

-0.87

รู้สึก

-0.83

POSITIVE LOGITS

use

1.63

用

1.14

 storing

1.09

 implement

1.07

使用

1.06

 checks

1.05

 store

1.05

 uses

1.03

 check

1.02

判断

1.01

Activations Density 0.667%