# Self-Consistency of the Internal Reward Models Improves Self-Rewarding Language Models

arXiv: https://arxiv.org/abs/2502.08922 · HTML: https://mlsanity.com/paper/2502.08922/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- Placed in Reinforcement learning for reasoning, one of 17 directions in LLM reasoning.
- Builds on 26 of its 36 references that are in the map.
- First posted on arXiv 13 Feb 2025.

## Placed in

- LLM reasoning / [Reinforcement learning for reasoning](https://mlsanity.com/lr/reinforcement-learning-for-reasoning/)
- LLM alignment / RLHF / preference optimisation / [LLM preference optimization methods](https://mlsanity.com/alignment/llm-preference-optimization-methods/)
- Synthetic data and self-improvement for LLMs / [Self-rewarding alignment and self-training](https://mlsanity.com/synth/self-rewarding-alignment-and-self-training/)

## Builds on (most cited first, 26 of 26)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [On the Opportunities and Risks of Foundation Models](https://mlsanity.com/paper/2108.07258/) (2021)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [Constitutional AI: Harmlessness from AI Feedback](https://mlsanity.com/paper/2212.08073/) (2022)
- [Unsupervised Data Augmentation for Consistency Training](https://mlsanity.com/paper/1904.12848/) (2019)
- [Length-Controlled AlpacaEval: A Simple Way to Debias Automatic Evaluators](https://mlsanity.com/paper/2404.04475/) (2024)
- [Discovering Latent Knowledge in Language Models Without Supervision](https://mlsanity.com/paper/2212.03827/) (2022)
- [Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2307.15217/) (2023)
- [OpenAssistant Conversations -- Democratizing Large Language Model Alignment](https://mlsanity.com/paper/2304.07327/) (2023)
- [Self-Rewarding Language Models](https://mlsanity.com/paper/2401.10020/) (2024)
- [Large Language Models Can Self-Improve](https://mlsanity.com/paper/2210.11610/) (2022)
- [Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision](https://mlsanity.com/paper/2312.09390/) (2023)
- [A Long Way to Go: Investigating Length Correlations in RLHF](https://mlsanity.com/paper/2310.03716/) (2023)
- [A Survey of Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2312.14925/) (2023)
- [Iterative Reasoning Preference Optimization](https://mlsanity.com/paper/2404.19733/) (2024)
- [Meta-Rewarding Language Models: Self-Improving Alignment with LLM-as-a-Meta-Judge](https://mlsanity.com/paper/2407.19594/) (2024)
- [Disentangling Length from Quality in Direct Preference Optimization](https://mlsanity.com/paper/2403.19159/) (2024)
- [Instruct2Act: Mapping Multi-modality Instructions to Robotic Actions with Large Language Model](https://mlsanity.com/paper/2305.11176/) (2023)
- [BioMedGPT: Open Multimodal Generative Pre-trained Transformer for BioMedicine](https://mlsanity.com/paper/2308.09442/) (2023)
- [Loose lips sink ships: Mitigating Length Bias in Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2310.05199/) (2023)
- [Consistency Regularization for Cross-Lingual Fine-Tuning](https://mlsanity.com/paper/2106.08226/) (2021)
- [CREAM: Consistency Regularized Self-Rewarding Language Models](https://mlsanity.com/paper/2410.12735/) (2024)

## Built on by (most cited first, 13 of 13)

- [Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models](https://mlsanity.com/paper/2503.09567/) (2025)
- [From System 1 to System 2: A Survey of Reasoning Large Language Models](https://mlsanity.com/paper/2502.17419/) (2025)
- [Consistent Paths Lead to Truth: Self-Rewarding Reinforcement Learning for LLM Reasoning](https://mlsanity.com/paper/2506.08745/) (2025)
- [Sentient Agent as a Judge: Evaluating Higher-Order Social Cognition in Large Language Models](https://mlsanity.com/paper/2505.02847/) (2025)
- [Can Reasoning Help Large Language Models Capture Human Annotator Disagreement?](https://mlsanity.com/paper/2506.19467/) (2025)
- [SPARTA ALIGNMENT: Collectively Aligning Multiple Language Models through Combat](https://mlsanity.com/paper/2506.04721/) (2025)
- [TraPO: A Semi-Supervised Reinforcement Learning Framework for Boosting LLM Reasoning](https://mlsanity.com/paper/2512.13106/) (2025)
- [Temporal Self-Rewarding Language Models: Decoupling Chosen-Rejected via Past-Future](https://mlsanity.com/paper/2508.06026/) (2025)
- [Can LLMs Learn to Reason Robustly under Noisy Supervision?](https://mlsanity.com/paper/2604.03993/) (2026)
- [LLMdoctor: Token-Level Flow-Guided Preference Optimization for Efficient Test-Time Alignment of Large Language Models](https://mlsanity.com/paper/2601.10416/) (2026)
- [Why Self-Rewarding Works: Theoretical Guarantees for Iterative Alignment of Language Models](https://mlsanity.com/paper/2601.22513/) (2026)
- [Seeing Beyond 8bits: Subjective and Objective Quality Assessment of HDR-UGC Videos](https://mlsanity.com/paper/2603.00938/) (2026)
- [J-Zero: Unified Challenger--Solver--Judge Self-Evolution from Zero Data](https://mlsanity.com/paper/2608.26582/) (2026)

## Abstract (arXiv)

Aligning Large Language Models (LLMs) with human preferences is crucial for their deployment in real-world applications. Recent advancements in Self-Rewarding Language Models suggest that an LLM can use its internal reward models (such as LLM-as-a-Judge) citeyuanself to generate preference data, improving alignment performance without costly human annotation. However, we find that different internal reward models within the same LLM often generate inconsistent preferences. This inconsistency raises concerns about the reliability of self-generated preference data, hinders overall alignment performance, and highlights the need for further research to ensure reliable and coherent alignment with human preferences. To address this limitation, we propose Self-Consistent Internal Rewards (SCIR), a novel framework designed to enhance consistency among internal reward models during training. In each training step, we collect preference predictions from multiple pre-defined internal reward models and enforce consistency and confidence through an inconsistency penalty mechanism, thereby improving the reliability of these internal reward models. We selectively use data with consistent predictions for preference optimization, ensuring the quality of the preference data. By employing self-consistent internal rewards, our method significantly improves the alignment performance and reward modeling capability of LLMs, outperforming baseline methods by a notable margin.