INDEX

Explanations

might be exploring alternatives

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

<unused2223>

1.48

 managable

1.21

ၣ်

1.20

 invincible

1.20

<unused2222>

1.20

inous

1.17

າມາດ

1.14

 COULD

1.13

ូប

1.13

<unused1157>

1.13

POSITIVE LOGITS

 inoltre

1.40

 asimismo

1.12

此外

1.11

 επίσης

1.10

However

1.09

Consequently

1.08

 natomiast

1.06

 pertanto

1.05

 berikutnya

1.04

 ardından

1.04

Activations Density 0.390%