INDEX

Explanations

precision and accuracy

The neuron detects occurrences of terms related to measurement precision or accuracy.

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

蟊

-1.59

 choses

-1.58

 mög

-1.58

跸

-1.55

 planètes

-1.52

 signes

-1.52

 EconPapers

-1.50

(«

-1.46

reasons

-1.44

 ruines

-1.44

POSITIVE LOGITS

1.73

 when

1.69

 accuracy

1.57

</b>

1.52

ot

1.42

 until

1.41

est

1.39

Activations Density 0.049%