INDEX

Explanations

methods, research, and frameworks

The neuron is picking out multi-digit numeric tokens—that is, large quantitative values (measurements, model coefficients, etc.) in technical or scientific contexts.

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

 ซึ่ง

1.16

していますが

1.06

がありますが

0.95

0.94

 whereby

0.94

؛

0.92

 nhưng

0.91

–

0.90

*,

0.89

 แต่

0.88

POSITIVE LOGITS

 futuras

0.84

 pudieran

0.82

 siquiera

0.78

 pudiera

0.75

 magari

0.74

 saurait

0.74

 pueda

0.74

போன்ற

0.74

 promoc

0.73

 futurs

0.73

Activations Density 0.050%