The ‘type’ parameter of your glsl array is set to ‘float’. Which mean it only takes the first channel of every sample. If you set this to vec4, it will take the first 4 channels. This is the maximum amount of channels to read into 1 variable (there is no such thing as a vec5, vec6 etc).
In your case I would use a shuffleCHOP to transform your chop into 1 channel with 10 samples. This way you can read them like lfo[0] to lfo[9] as a float.