INDEX

Explanations

normalization and standardization

This neuron activates on tokens that represent numbers or numeric expressions (i.e., any token containing digits).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 santa

0.56

newtheorem

0.55

 خون

0.55

 Tsuen

0.54

 एलान

0.54

 coscienza

0.54

 Isolated

0.54

 spiaggia

0.54

santa

0.54

McM

0.52

POSITIVE LOGITS

[:,

1.08

 matrix

1.02

matrix

0.98

Matrix

0.95

 Matrix

0.93

 matrices

0.89

矩阵

0.87

 columns

0.86

(:,:,

0.84

 MATRIX

0.82

Activations Density 0.020%