INDEX

Explanations

in phrases specifying parts

The neuron strongly activates on numeric tokens with decimal points (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

這

0.65

 Kenyon

0.61

 beforehand

0.58

。

0.57

此

0.56

 Publ

0.55

 Derecho

0.55

 切り

0.54

ón

0.54

 Daarna

0.54

POSITIVE LOGITS

すなわち

0.81

 включая

0.73

orod

0.71

 जिसमें

0.66

 incluindo

0.66

รวม

0.66

 குறிப்பாக

0.66

あるいは

0.64

जिस

0.63

 зокрема

0.63

Activations Density 0.074%