INDEX

Explanations

numbers or round indicators

The neuron strongly lights up on numeric and statistical tokens—years, scores, counts, and the punctuation around them.

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

 মানুষের

0.88

 расти

0.84

 plasticity

0.84

 aesthetic

0.83

 κοινων

0.83

植物

0.82

 postural

0.82

삶

0.82

 الإنسان

0.82

 ადამიან

0.82

POSITIVE LOGITS

 three

1.31

 five

1.27

 four

1.25

six

1.25

 trois

1.23

 तीन

1.22

 quatro

1.19

 seven

1.17

three

1.17

 drei

1.14

Activations Density 1.223%