INDEX

Explanations

git clone

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 terbesar

1.09

 Rashid

1.07

 Mohd

1.06

 rhinoceros

1.04

こうした

1.02

/_

1.01

 lainnya

1.00

 dichos

0.99

Muh

0.98

 Pelham

0.98

POSITIVE LOGITS

 desirable

1.11

 dynamically

1.07

编译器

1.02

 истори

1.01

 allowed

0.99

 argue

0.98

 程序

0.98

约束

0.96

 bounds

0.95

CHANT

0.94

Activations Density 0.005%

No Known Activations