Points of (0,1) are trivially in the closure. Let x=0: for every ε>0, the witness min(ε/2,1/2)∈(0,ε)∩(0,1) (the min is needed since ε/2 can leave (0,1) when ε≥2), so every neighborhood of 0 meets (0,1); i.e. 0∈(0,1).
Likewise every ε-ball at 1 contains 1−min(ε/2,1/2)∈(0,1), so 1∈(0,1).
If x<0, the ball of radius −x/2 around x lies in (−∞,0), disjoint from (0,1); similarly for x>1. Hence the closure is exactly [0,1].