INDEX

Explanations

dates and years

The neuron fires on numeric data tokens—years, percentages, counts, dates—i.e. any explicit number or date reference in the text.

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

昨日の

-1.37

Different

-1.35

 berbeda

-1.26

先日

-1.23

different

-1.20

Caratteristiche

-1.16

ikations

-1.14

 diferite

-1.14

 comuniques

-1.13

توضیحات

-1.13

POSITIVE LOGITS

 October

1.44

we

1.41

 June

1.38

 August

1.32

 April

1.31

 January

1.30

 March

1.26

May

1.24

 February

1.23

 November

1.22

Activations Density 0.024%