The key problem here is that we don’t know what rewards we “would have” provided in situations that did not occur during training. This requires us to choose some specific counterfactual, to define what “would have” happened. After we choose a counterfactual, we can then categorize a failure as outer or inner misalignment in a well-defined manner.
We often do know what rewards we “would have” provided. You can query the reward function, reward model, or human labellers. IMO, the key issue with the objective-based categorisation is a bit different: it’s nonsensical to classify an alignment failure as inner/outer based on some value of the reward function in some situation that didn’t appear during training, as that value has no influence on the final model.
In other words: Maybe we know what reward we “would have” provided in a situation that did not occur during training, or maybe we don’t. Either way, this hypothetical reward has no causal influence on the final model, so it’s silly to use this reward to categorise any alignment failures that show up in the final model.
We often do know what rewards we “would have” provided. You can query the reward function, reward model, or human labellers. IMO, the key issue with the objective-based categorisation is a bit different: it’s nonsensical to classify an alignment failure as inner/outer based on some value of the reward function in some situation that didn’t appear during training, as that value has no influence on the final model.
In other words: Maybe we know what reward we “would have” provided in a situation that did not occur during training, or maybe we don’t. Either way, this hypothetical reward has no causal influence on the final model, so it’s silly to use this reward to categorise any alignment failures that show up in the final model.
Yeah, that’s another good reason to be skeptical of the objective-based categorization.