INDEX

Explanations

start of sentence or name

The neuron is effectively inactive (a “dead” neuron) – it does not reliably fire for any patterns in the input.

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

ing

-2.41

 With

-2.22

are

-1.97

 Another

-1.91

十分钟

-1.82

最后由

-1.81

 While

-1.77

 would

-1.77

we

-1.75

ጪ

-1.72

POSITIVE LOGITS

ه

2.41

 ſol

2.09

因为

2.02

为了

1.87

她的

1.73

让

1.72

 ürünleri

1.68

Характеристики

1.66

1.65

 eſſ

1.63

Activations Density 0.000%