That's not the goal of a VM/runtime, though: it only ensures that the hosted code cannot escape the sandbox, and provides a correct implementation of the semantics of each VM opcode. The hosted code's computation implemented with those opcodes is a black box to the VM.

Consider: should the VM somehow try to analyze the running code and determine when it's about to commit a logical error and return "forbidden" data? What specification states which data is forbidden?

Consider: even the most high-level VM can execute a program with logical errors if it is Turing-complete. A Java or Erlang program running on a bug-free, fully-compliant implementation of the respective VM can still get (e.g.) an authorization condition wrong, or incorrectly implement a security protocol, or return data from the wrong (in-bounds, but wrong!) index in an array.