INDEX

Explanations

device functions

The neuron strongly responds to occurrences of the term “function” (and closely related feature‐describing words) in technical/patent‐style descriptions.

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

戕

-1.57

 there

-1.55

 favorecer

-1.52

daad

-1.52

 aliviar

-1.44

ेमाल

-1.41

 empuj

-1.39

 recaud

-1.38

salmon

-1.38

 スクール

-1.37

POSITIVE LOGITS

2.00

1.94

1.87

es

1.74

1.66

se

1.61

1.57

1.56

但却

1.56

其

1.55

Activations Density 0.063%