INDEX

Explanations

logical deduction

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

small

1.25

inner

1.18

oven

1.15

ch

1.07

previously

1.07

ịn

1.06

 insides

1.04

 small

1.04

皮

1.02

chairs

0.99

POSITIVE LOGITS

 regola

1.46

 вероят

1.42

 assicur

1.41

 molte

1.41

লালের

1.35

 Percolation

1.34

 conseguenza

1.34

 Atomnoj

1.34

 McKinsey

1.32

িঝ

1.32

Activations Density 0.005%