| Factor | Self-Consistency | Iterative Refinement |
|---|---|---|
| Approach | Multiple parallel generations | Sequential improvements |
| Selection Method | Majority voting/consistency | Feedback-based adjustment |
| Iterations | Single round (parallel) | Multiple rounds (sequential) |
| Cost | High (multiple completions) | Variable (depends on iterations) |
| Use Case | Reasoning tasks | Quality improvement |
| Human Involvement | Minimal | Can be significant |
| Convergence | Immediate | Gradual |
| Best For | Reducing variance | Achieving specific quality |
Use Self-Consistency Methods when you need to improve reliability on reasoning tasks where multiple valid solution paths exist, when you want to reduce the impact of random variation in model outputs, when you can afford multiple API calls per query, when the task has a clear correct answer that can be verified through agreement, or when you need confidence estimates based on response consistency. It's ideal for math problems, logical reasoning, question answering, and scenarios where the cost of errors is high and computational cost is acceptable.
Use Iterative Refinement when you need to progressively improve output quality toward specific criteria, when initial outputs are close but not quite right, when you have clear feedback mechanisms (human or automated), when you're optimizing for subjective quality dimensions, or when you want to incorporate learning from previous attempts. It's essential for creative tasks, content generation, prompt development itself, complex writing assignments, and scenarios where quality requirements are nuanced and may require multiple attempts to satisfy.
Combine both by using Self-Consistency to generate multiple candidate outputs, then apply Iterative Refinement to the most consistent or promising candidate to polish it further. For example, generate 5 solutions using self-consistency to identify the most likely correct approach, then iteratively refine that solution to improve clarity, completeness, or style. You can also use self-consistency at each iteration of refinement to ensure improvements are consistent. This combination provides both reliability (from self-consistency) and quality optimization (from refinement).
Self-Consistency generates multiple independent outputs in parallel and selects the best through voting or agreement, focusing on finding the most reliable answer among variations. Iterative Refinement generates one output, evaluates it, provides feedback, and generates an improved version sequentially, focusing on progressive quality improvement. Self-consistency is about sampling diversity to find consensus; refinement is about directed improvement toward a goal. Self-consistency requires multiple simultaneous API calls; refinement requires sequential calls with feedback loops. Self-consistency works best when there's a 'correct' answer; refinement works best when quality is subjective or multidimensional.
Many believe self-consistency always improves results, but it only helps when multiple reasoning paths lead to the same answer—it can't fix fundamental model limitations. Others think iterative refinement always converges to better outputs, but without good feedback, it can plateau or even degrade. Some assume self-consistency is just 'running the prompt multiple times,' missing the importance of the aggregation method. Another misconception is that refinement requires human feedback, when automated evaluation can drive it. Finally, users often underestimate the cost of self-consistency, which multiplies API calls by the number of samples.