INDEX

Explanations

number of

This neuron detects explicit numeric counts, i.e. digits appearing after words like “number” to indicate how many of something there are.

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

-2.41

la

-2.08

瑷

-2.03

CurtirCurtir

-1.88

 可愛い

-1.85

消费者

-1.84

我怎么

-1.82

Ս

-1.81

^{*},

-1.80

𖤣

-1.80

POSITIVE LOGITS

),

2.38

of

2.27

2.13

先日

1.94

''

1.87

 bekom

1.80

Along

1.76

1.73

 kurzem

1.72

 perfekte

1.70

Activations Density 0.022%