INDEX

Explanations

earlier, previous, last

The neuron flags phrases that talk about one event following the same format or structure as previous editions.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 Gewerks

-0.87

*)(

-0.85

cambi

-0.84

OUT

-0.84

and

-0.84

Mvh

-0.83

 LoginComponent

-0.82

KY

-0.81

Когда

-0.81

 když

-0.81

POSITIVE LOGITS

 previous

3.39

 last

3.22

 предыду

2.42

 previously

2.09

 earlier

2.09

previous

2.08

 sebelumnya

1.97

 poprzed

1.95

 recent

1.91

 précédente

1.85

Activations Density 0.077%