INDEX

Explanations

contrasting alternatives

The neuron fires on numeric tokens (both integers and decimals) appearing 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

 không

-1.35

metries

-1.27

 يول

-1.20

を見る

-1.20

of

-1.19

自ら

-1.15

太多了

-1.15

⅘

-1.13

又不

-1.12

 existência

-1.11

POSITIVE LOGITS

 than

1.60

、「

1.41

 нього

1.35

 mainly

1.30

 some

1.24

 Instead

1.23

 活動

1.20

or

1.20

 Makes

1.20

 which

1.20

Activations Density 0.036%