Elimination of Bias in Introspection: Methodological Advances, Refinements, and Recommendations — LessWrong