INDEX

Explanations

breaking down into smaller steps

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

ភាព

1.02

ostat

0.98

the

0.95

Polit

0.91

जिक

0.89

 fossa

0.88

ванием

0.88

ensen

0.88

 imaju

0.86

ax

0.86

POSITIVE LOGITS

 vassals

1.07

 Modules

1.03

 PARTS

1.03

 clumps

1.02

 tarea

1.01

 Agents

1.01

 dividido

1.01

 subunits

1.00

 Apprentice

0.99

 replicas

0.99

Activations Density 0.091%