Large Language Models are Better Reasoners with Self-Verification

By Yixuan Weng, Minjun Zhu, Fei Xia, Bin Li, Shizhu He, Kang Liu, Jun Zhao. Conference on Empirical Methods in Natural Language Processing, 2022

424 citations recorded by the source corpus.

Recently, with the chain of thought (CoT) prompting, large language models (LLMs), e.g., GPT-3, have shown strong reasoning ability in several natural language processing tasks such as arithmetic, commonsense, and logical reasoning. However, LLMs with CoT require multi-step prompting and multi-token prediction, which is highly sensitive to individual mistakes and vulnerable to error accumulation. The above issues make the LLMs need the ability to verify the answers. In fact, after inferring conclusions in some thinking decision tasks, people often check them by re-verifying steps to avoid some mistakes. In this paper, we propose and prove that LLMs also have similar self-verification abilities. We take the conclusion obtained by CoT as one of the conditions for solving the original problem. By performing a backward verification of the answers that LLM deduced for itself, we can obtain interpretable answer validation scores to select the candidate answer with the highest score. Experimental results demonstrate that the proposed method can improve the reasoning performance on various arithmetic, commonsense, and logical reasoning datasets. Our code is publicly available at: https://github.com/WENGSYX/Self-Verification.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/7715ba5e75f5256e1061c7473afe61bb0dbb9065

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

LlaVA-CoT: Let Vision Language Models Reason Step-By-Step (2024, neutral): “Existing methods for inference time scaling fall into two main categories: those that rely on an external verifier for selection [27, 60] and those that operate independently of any external verifier [21, 57].”
R1-Onevision: Advancing Generalized Multimodal Reasoning Through Cross-Modal Formalization (2025, neutral): “For example, Chain-of-Thought (CoT) prompting, where a complex question is decomposed into intermediate reasoning steps, has shown promise in guiding LLMs to structured solutions [29, 34].”
LoRA Learns Less and Forgets Less (2024, methodology): “Both Self-Verification (Weng et al., 2022) and FOBAR (Jiang et al., 2024b) fall under the category of “backward reasoning,” where the question starts with a given condition and requires reasoning backwards to solve for an unknown variable.”
The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity (2025, positive): “For instance, generating a Chain of Thought (CoT) [15, 16, 17, 18] and incorporating self-verification [19, 20, 21] prior to the final answer have been shown to improve model performance.”
LLM Self Defense: By Self Examination, LLMs Know They Are Being Tricked (2023, background): “This falls in line with their ability to verify the content they generate for accuracy and then correct themselves using a chain of verification, which has been shown to improve factual question answering data [6], mathematical reasoning [19], and logical reasoning [29].”
MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Models (2023, methodology): “For the backward direction, we have the self-verification question [76] and FOBAR question [32].”
Why Do Multi-Agent LLM Systems Fail? (2025, neutral): “Verification varies by domain: coding requires thorough test coverage, QA demands certified data checks (Peng et al., 2023), and reasoning benefits from symbolic validation (Kapanipathi et al., 2020).”
Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning (2025, positive): “With RL, even when trained solely on the outcome rewards, the models learn complex reasoning capabilities, including self-verification (Weng et al., 2022) and self-correction (Kumar et al., 2024).”
Syntriever: How to Train Your Retriever with Synthetic Data from LLMs (2025, unknown): “LLMs are known to have self-verification ability (Weng et al., 2023), i.e., they can re-verify the inferred answer.”