INDEX

Explanations

series numbers

The neuron fires on occurrences of the word “Series” (or its variants) paired with a following numeral, i.e. “Series X” labels.

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

eronau

-1.90

暇

-1.86

 McInt

-1.80



-1.79

 sleek

-1.77

spectos

-1.76

вера

-1.75

 ersche

-1.74

 slated

-1.74

 terza

-1.73

POSITIVE LOGITS

.”

2.83

 рисун

1.98

 behöver

1.88

of

1.80

 olika

1.80

桝

1.79

非常に

1.75

の味

1.74

</h3>

1.73

 Series

1.73

Activations Density 0.027%