INDEX

Explanations

code programming keywords

The neuron is strongly activated by numeric tokens (numbers and numeric identifiers).

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

canic

0.54

lgende

0.52

¹)

0.51

Lambda

0.50

kV

0.50

ti

0.50

 केले

0.50

izontal

0.49

 නිෂ්

0.48

 unpacked

0.48

POSITIVE LOGITS

Waq

0.62

 isValid

0.56

 এখনো

0.55

 miembro

0.54

 Badal

0.53

 Still

0.52

 ยัง

0.52

 member

0.51

 parody

0.50

 Bade

0.50

Activations Density 0.001%