INDEX

Explanations

placeholders and context

The neuron is triggered by numeric tokens representing measured or statistical values (i.e. floating-point numbers).

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

 यही

1.12

 사람

0.93

The

0.91

 Также

0.88

 They

0.87

It

0.86

 Этот

0.85

 Οι

0.84

 Это

0.84

ელის

0.84

POSITIVE LOGITS

↵↵

1.45

↵

1.27

↵↵↵

1.11

↵↵↵↵

0.92

‬

0.88

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.87

↵↵↵↵↵

0.86

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.85

</h2>

0.83

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.80

Activations Density 0.836%