Internal rationality is using truth-preserving thought processes that output good hypotheses given good data. Examples include:

• Identifying inconsistencies in hypotheses.
• Having detailed models/inside views.
• Not doing motivated reasoning.
• Evaluating one's thought processes.

External rationality is discerning good data from bad data. Examples include:

• Identifying the selection effects and incentives behind what data one sees.
• Identifying whether some data is "adversarial": data that strongly suggests a hypothesis just because there will inevitably be one source of data that strongly suggests a given hypothesis just because there are so many sources of data.
• Knowing when to use an outside view.
• Knowing what experts to trust on a topic one knows nothing about.

External rationality and internal rationality intersect and build upon one another.

I like using these words instead of "inside view" and "outside view" because "outside view" tends to exclude model-based reasoning when you can have detailed models of what data is good or bad. Also "outside view" specifically refers to reference class forecasting and I don't want to overload the term. I also feel like "external rationality" and "internal rationality" are better at communicating the fuzzy intuition I'm getting at here.