# CAP-CoT: Cycle Adversarial Prompt for Improving Chain of Thoughts in LLM Reasoning

arXiv: https://arxiv.org/abs/2604.23270 · HTML: https://mlsanity.com/paper/2604.23270/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning.
- Builds on 12 of its 57 references that are in the map.
- First posted on arXiv 25 Apr 2026.

## Placed in

- LLM reasoning / [Prompting and chain-of-thought reasoning](https://mlsanity.com/lr/prompting-and-chain-of-thought-reasoning/)
- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 12 of 12)

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Prompt Injection attack against LLM-integrated Applications](https://mlsanity.com/paper/2306.05499/) (2023)
- [Contrastive Chain-of-Thought Prompting](https://mlsanity.com/paper/2311.09277/) (2023)
- [Think-While-Generating: On-the-Fly Reasoning for Personalized Long-Form Generation](https://mlsanity.com/paper/2512.06690/) (2025)
- [Look Inward to Explore Outward: Learning Temperature Policy from LLM Internal States via Hierarchical RL](https://mlsanity.com/paper/2602.13035/) (2026)
- [Lightweight LLM Agent Memory with Small Language Models](https://mlsanity.com/paper/2604.07798/) (2026)
- [LLaVA-FA: Learning Fourier Approximation for Compressing Large Multimodal Models](https://mlsanity.com/paper/2602.00135/) (2026)
- [When Right Meets Wrong: Bilateral Context Conditioning with Reward-Confidence Correction for GRPO](https://mlsanity.com/paper/2603.13134/) (2026)
- [Experience Transfer for Multimodal LLM Agents in Minecraft Game](https://mlsanity.com/paper/2604.05533/) (2026)
- [Your Language Model May Think Too Rigidly: Achieving Reasoning Consistency with Symmetry-Enhanced Training](https://mlsanity.com/paper/2502.17800/) (2025)
- [TDA-RC: Task-Driven Alignment for Knowledge-Based Reasoning Chains in Large Language Models](https://mlsanity.com/paper/2604.04942/) (2026)
- [Transforming External Knowledge into Triplets for Enhanced Retrieval in RAG of LLMs](https://mlsanity.com/paper/2604.12610/) (2026)

## Built on by (most cited first, 1 of 1)

- [From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AI](https://mlsanity.com/paper/2606.14502/) (2026)

## Abstract (arXiv)

Chain-of-Thought (CoT) prompting has emerged as a simple and effective way to elicit step-by-step solutions from large language models (LLMs). However, CoT reasoning can be unstable across runs on long, multi-step problems, leading to inconsistent answers for unchanged task. Most prior work focuses on improving the forward reasoning chain within a single pass, with less attention to iterative and contrastive correction. To address this gap, we propose CAP-CoT, a Cycle Adversarial Prompt optimization framework designed to improve both CoT reasoning accuracy and stability of a single deployed solver. In each cycle, a forward solver generates candidate reasoning chains, an adversarial challenger constructs plausible but deliberately flawed chains using targeted error strategies, and a feedback agent contrasts the two chains and produces step-aligned structured feedback. This feedback closes the optimization loop in two directions, including updating the solver prompt based on errors exposed by the challenger, and updating the challenger prompt to generate increasingly targeted errors in subsequent cycles. Unlike safety-oriented adversarial prompting such as jailbreak or prompt-injection attacks, our adversarial component is task-semantic and aims to expose logical vulnerabilities in reasoning chains. Experiments across six benchmarks and four LLM backbones demonstrate that within two to three adversarial prompt optimization cycles, CAP-CoT consistently reduces variability across runs while improving reasoning accuracy and robustness to prompt perturbations.