A large language model was asked to solve the following classic word problem: 'A bat and a ball cost $1.10 in total. The bat costs $1.00 more than the ball. How much does the ball cost?' The model initially and incorrectly answered: 'The ball costs $0.10.'

Your task is to apply one cycle of an iterative improvement framework. Describe the next two distinct stages of this process. For each stage, explain its purpose and provide a specific, well-formed prompt you would use to guide the model toward the correct answer.

Google

Self-refinement is an iterative framework where a solution is progressively improved through repeated cycles of generation, evaluation (critique), and revision. This process continues until the solution's quality, as assessed by a verifier, converges to a satisfactory level.

Self-Refinement in LLMs

A practical application of self-refinement involves a two-step process for machine translation. First, a Large Language Model is prompted to translate a text, for instance from Chinese to English. Subsequently, the same model is prompted again to refine its initial translation, aiming to improve its quality and accuracy.

Example of Self-Refinement in Machine Translation

A widely recognized general framework for self-refinement using Large Language Models consists of a three-step process, as outlined by Madaan et al. (2024).

Three-Step Framework for Self-Refinement in LLMs

In an ideal scenario, a highly capable Large Language Model would be able to execute all three steps of the self-refinement process—prediction, feedback, and revision—effectively without requiring any additional, specialized training.

Ideal Self-Refinement without Additional Training

When sufficient labeled data is available for a specific task, the self-refinement capabilities of a Large Language Model can be improved through supervised learning. A common approach is to fine-tune the LLM, which adapts it to better handle refinement processes.

Fine-Tuning LLMs for Self-Refinement Tasks

As an alternative to fine-tuning a general-purpose Large Language Model for refinement, specialized, task-specific models can be developed using supervised learning when sufficient labeled data is available. These models, which are not necessarily based on LLM architectures, provide another method for enhancing performance on specific refinement tasks.

Task-Specific Models as an Alternative for Refinement

The challenge of improving the self-refinement capabilities of Large Language Models can be framed as an alignment problem. This perspective considers the process of enhancing self-correction and refinement as a way of guiding the model's behavior to be more consistent with desired outcomes and human intentions.

Self-Refinement as an LLM Alignment Issue

Self-reflection in Large Language Models is a concept analogous to human introspection, where the model evaluates its own outputs. It is believed that if LLMs can self-reflect, they can achieve greater accuracy and develop self-correction capabilities, thereby improving their predictions.

Self-Reflection in LLMs

A developer is using a large language model to generate a Python function for a complex data analysis task. The developer's workflow is as follows:

1. The model generates an initial version of the function.
2. The developer then prompts the same model, providing the initial function and asking it to 'act as a senior code reviewer, identify potential bugs or inefficiencies, and explain how to fix them.'
3. Based on the model's feedback, a final, improved version of the function is produced.

This iterative process of generating an output, using the model to critique its own output, and then improving it based on that critique is best described as:

Applying an Iterative Improvement Framework

The process of self-refinement in AI can be compared to the iterative design of a product. A designer often starts with a basic prototype, which is then evaluated and tested. Based on this feedback, the design is refined to improve its functionality and user experience. This cycle of creation, evaluation, and refinement may be repeated multiple times until the product meets the desired standards.

Product Design as an Analogy for Self-Refinement

Self-refinement in Large Language Models is closely associated with concepts that touch upon the psychological aspects of these models, with self-reflection being a prominent example. The capacity for a model to self-reflect is considered a foundational element that underpins its ability to self-correct and improve its predictions.

Relationship between Self-Refinement and Self-Reflection in LLMs

Explain the fundamental difference between a process where a model generates a single output and then iteratively improves it, versus a process where a model generates multiple different outputs at once and then selects the best one. In your explanation, characterize the primary goal of each approach.

Comparing Output Improvement Strategies

Your team is rolling out an internal LLM assistant...

You’re building an internal LLM workflow to produc...

You’re building an internal LLM assistant to help ...

You’re leading an internal enablement team buildin...

You are deploying an LLM inside an internal finance tool that must answer multi-step policy questions (e.g., “Is this expense reimbursable given these 6 policy clauses and these 3 trip details?”). Constraints: (1) you have a strict context window budget, so you can include at most ONE short worked example in the prompt; (2) the model’s first-pass answers are often incomplete (it sometimes explains reasoning but forgets to state a final decision); (3) you need a workflow that is reliable across many different question shapes, not just one template.

Write an essay proposing a concrete prompting workflow that combines: in-context learning (if you use it), chain-of-thought prompting (including whether you would use zero-shot CoT), explicit problem decomposition (including whether you would use a least-to-most progression), and an iterative self-refinement loop. Your answer must (a) explain why your chosen combination fits the constraints, (b) describe the exact sequence of LLM calls or prompt sections you would use, (c) explain how intermediate sub-answers feed into later steps, and (d) specify how you will force a clear final decision despite the risk of incomplete zero-shot CoT outputs. Include at least two tradeoffs you are accepting (e.g., latency vs. accuracy, context budget vs. guidance, error propagation vs. transparency) and how you would mitigate them.

Choosing and Justifying a Prompting Strategy Under Context and Quality Constraints

You are deploying an LLM inside a corporate workflow to draft a first-pass analysis for a high-stakes internal decision memo (e.g., whether to approve a vendor contract). The input is a long, messy bundle of notes and constraints, and early pilots show two failure modes: (1) when asked directly, the model misses key constraints and produces confident but shallow conclusions; (2) when you add a generic “let’s think step by step” instruction, it produces lengthy reasoning but sometimes never states a clear final recommendation.

Write a proposed prompting workflow (you may describe it as a sequence of prompts) that uses: in-context learning with demonstrations, chain-of-thought prompting (including a zero-shot trigger where appropriate), least-to-most prompting, explicit problem decomposition, and an iterative self-refinement loop. Your answer must explain how these techniques work together (not independently) to reduce the two failure modes above, and must justify at least two tradeoffs you are making (e.g., latency/cost vs. accuracy, transparency vs. risk of overlong reasoning, fixed vs. dynamic decomposition). Conclude by giving one concrete example of a “final answer formatting” instruction you would include to ensure the model always outputs a clear recommendation at the end.

Designing a Prompting Workflow for a High-Stakes, Multi-Step Task

You are building an internal LLM assistant that helps a finance operations team investigate invoice-payment discrepancies. A typical ticket requires (a) extracting key entities from a messy email thread, (b) reconciling amounts across two systems, (c) identifying the most likely root cause (e.g., duplicate invoice, partial payment, currency conversion, timing), and (d) drafting a customer-facing explanation.

Your first implementation uses a single prompt with the full ticket text and the instruction “Let’s think step by step and give the final answer.” It often produces plausible reasoning but (1) misses a critical sub-check (e.g., currency conversion), (2) sometimes never states a clear final conclusion, and (3) varies in format so downstream automation breaks.

Write an essay that proposes a revised prompting workflow that integrates: in-context learning (demonstrations), explicit problem decomposition, least-to-most sequencing of sub-problems, chain-of-thought prompting (including when you would use zero-shot CoT vs a demonstrated CoT example), and an iterative self-refinement loop. Your answer must explain how these elements work together (not as independent tips), what tradeoffs you are making (e.g., context length, reliability, error propagation), and include at least two concrete prompt snippets (not full prompts) that illustrate (i) how you would elicit structured intermediate reasoning while still forcing a final answer, and (ii) how you would run critique-and-revision in self-refinement without losing the decomposed structure.

Diagnosing and Redesigning a Prompting Approach for a Decomposed Workflow

You are deploying an internal LLM assistant to help procurement analysts decide whether a proposed vendor contract requires (a) a standard review, (b) an enhanced review, or (c) an automatic rejection. The decision depends on multiple interdependent rules (e.g., data types handled, cross-border transfers, subcontractors, and exception clauses). In pilot testing, a single prompt that asks for the final decision often misses a key condition; a zero-shot “Let’s think step by step” prompt sometimes produces a long rationale but forgets to clearly state the final decision; and when you add a few demonstrations, the model becomes more consistent but still makes occasional early-step mistakes that cascade into the wrong outcome.

Design a prompting workflow (you may describe it as a sequence of LLM calls) that uses: (1) in-context learning via demonstrations, (2) explicit problem decomposition in a least-to-most progression, and (3) an iterative self-refinement loop. Your answer must explain how information flows from one step to the next, how you would prevent or detect cascading errors from early steps, and how you would ensure the model always outputs an unambiguous final decision label (a/b/c) even when using chain-of-thought style reasoning.

Stabilizing an LLM Workflow for Multi-Step Policy Compliance Decisions

You are rolling out an internal LLM assistant for Legal Ops to triage vendor contract clauses into three labels: (A) "standard/low risk", (B) "needs legal review", (C) "reject". The assistant must also produce a short justification that can be audited. In a pilot, you observe two recurring failures: (1) on complex clauses, the model gives confident but wrong labels unless it is guided through intermediate reasoning; (2) when you add a simple instruction like "Let’s think step by step," the model often produces a long rationale but sometimes forgets to output a clear final label.

You are not allowed to fine-tune the model. You can only change the prompting workflow and the content placed in the prompt. You have a strict context window budget, so you can include at most ONE worked example in the prompt, and you must keep the number of model calls per clause to no more than 3.

Case: A new clause says: "Vendor may use Customer Data to improve its services and may share Customer Data with affiliates and subcontractors for that purpose. Customer may opt out of data sharing by emailing support within 10 days of signing." Your policy requires: (i) no sharing of Customer Data with third parties for model training/service improvement unless explicitly prohibited or strongly constrained; (ii) opt-out mechanisms are considered insufficient for sensitive data; (iii) subcontractor access must be tightly limited.

Design a prompting workflow (describe the sequence of prompts/calls and what each prompt contains) that uses in-context learning, problem decomposition/least-to-most prompting, chain-of-thought (including zero-shot CoT where appropriate), and self-refinement to reduce both failure modes while staying within the constraints. Your answer must explain how the techniques interact (e.g., why you place a worked example where you do, how you decompose the decision, how you ensure a final label is always produced, and how critique/revision is used without exceeding 3 calls).

Debugging a Multi-Step LLM Workflow for Contract Clause Risk Triage

You are rolling out an internal LLM assistant for SREs to perform root-cause analysis (RCA) from incident artifacts. The assistant receives: (1) a short incident timeline, (2) 30–60 lines of logs, and (3) a list of recent deploys. The desired output is a single RCA summary with: likely cause, supporting evidence, and a recommended next action.

Constraints: You can only include ONE short demonstration in the prompt due to context limits; the assistant must work on novel incidents; and the first pilot showed two failure modes: (a) the model jumps to a plausible cause without checking evidence across artifacts, and (b) when you add “Let’s think step by step,” it sometimes produces a long analysis but forgets to state a clear final RCA.

As the prompt/workflow designer, propose a concrete prompting plan that combines: (i) in-context learning with the single demonstration, (ii) explicit problem decomposition (including a least-to-most style progression of sub-questions), (iii) chain-of-thought elicitation (including how you would use or avoid zero-shot CoT triggers), and (iv) self-refinement (critique + revision) to reduce both failure modes. Your answer must include the exact structure of the prompt(s) you would send (section headings and what each section contains) and explain why each component is placed where it is.

Learn Before

Related