# BPO: Towards Balanced Preference Optimization between Knowledge Breadth and Depth in Alignment

arXiv: https://arxiv.org/abs/2411.10914 · HTML: https://mlsanity.com/paper/2411.10914/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- Placed in LLM preference optimization methods, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 30 of its 45 references that are in the map.
- First posted on arXiv 16 Nov 2024.

## Placed in

- LLM alignment / RLHF / preference optimisation / [LLM preference optimization methods](https://mlsanity.com/alignment/llm-preference-optimization-methods/)

## Builds on (most cited first, 30 of 30)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [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)
- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [LIMA: Less Is More for Alignment](https://mlsanity.com/paper/2305.11206/) (2023)
- [SimPO: Simple Preference Optimization with a Reference-Free Reward](https://mlsanity.com/paper/2405.14734/) (2024)
- [Length-Controlled AlpacaEval: A Simple Way to Debias Automatic Evaluators](https://mlsanity.com/paper/2404.04475/) (2024)
- [Safe RLHF: Safe Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2310.12773/) (2023)
- [Estimating Training Data Influence by Tracing Gradient Descent](https://mlsanity.com/paper/2002.08484/) (2020)
- [LESS: Selecting Influential Data for Targeted Instruction Tuning](https://mlsanity.com/paper/2402.04333/) (2024)
- [From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipeline](https://mlsanity.com/paper/2406.11939/) (2024)
- [What Makes Good Data for Alignment? A Comprehensive Study of Automatic Data Selection in Instruction Tuning](https://mlsanity.com/paper/2312.15685/) (2023)
- [Statistical Rejection Sampling Improves Preference Optimization](https://mlsanity.com/paper/2309.06657/) (2023)
- [A Survey on Data Selection for Language Models](https://mlsanity.com/paper/2402.16827/) (2024)
- [Benchmarking Foundation Models with Language-Model-as-an-Examiner](https://mlsanity.com/paper/2306.04181/) (2023)
- [Human-like Summarization Evaluation with ChatGPT](https://mlsanity.com/paper/2304.02554/) (2023)
- [Calibrating Sequence likelihood Improves Conditional Language Generation](https://mlsanity.com/paper/2210.00045/) (2022)
- [MM-LIMA: Less Is More for Alignment in Multi-Modal Datasets](https://mlsanity.com/paper/2308.12067/) (2023)
- [DALK: Dynamic Co-Augmentation of LLMs and KG to answer Alzheimer's Disease Questions with Scientific Literature](https://mlsanity.com/paper/2405.04819/) (2024)
- [Fine-Tuning Language Models with Advantage-Induced Policy Alignment](https://mlsanity.com/paper/2306.02231/) (2023)
- [Understanding In-Context Learning via Supportive Pretraining Data](https://mlsanity.com/paper/2306.15091/) (2023)
- [Curry-DPO: Enhancing Alignment using Curriculum Learning & Ranked Preferences](https://mlsanity.com/paper/2403.07230/) (2024)
- [Insights into Alignment: Evaluating DPO and its Variants Across Multiple Tasks](https://mlsanity.com/paper/2404.14723/) (2024)
- [Scaling Data Diversity for Fine-Tuning Language Models in Human Alignment](https://mlsanity.com/paper/2403.11124/) (2024)
- [Self-Training with Direct Preference Optimization Improves Chain-of-Thought Reasoning](https://mlsanity.com/paper/2407.18248/) (2024)

## Built on by (most cited first, 20 of 20)

- [From Generation to Judgment: Opportunities and Challenges of LLM-as-a-judge](https://mlsanity.com/paper/2411.16594/) (2024)
- [Preference Leakage: A Contamination Problem in LLM-as-a-judge](https://mlsanity.com/paper/2502.01534/) (2025)
- [TreeReview: A Dynamic Tree of Questions Framework for Deep and Efficient LLM-based Scientific Peer Review](https://mlsanity.com/paper/2506.07642/) (2025)
- [Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models](https://mlsanity.com/paper/2601.14004/) (2026)
- [Understanding and Mitigating Overrefusal in LLMs from an Unveiling Perspective of Safety Decision Boundary](https://mlsanity.com/paper/2505.18325/) (2025)
- [TL;DR: Too Long, Do Re-weighting for Efficient LLM Reasoning Compression](https://mlsanity.com/paper/2506.02678/) (2025)
- [Find Your Optimal Teacher: Personalized Data Synthesis via Router-Guided Multi-Teacher Distillation](https://mlsanity.com/paper/2510.10925/) (2025)
- [ToolPRMBench: Evaluating and Advancing Process Reward Models for Tool-using Agents](https://mlsanity.com/paper/2601.12294/) (2026)
- [Larger or Smaller Reward Margins to Select Preferences for Alignment?](https://mlsanity.com/paper/2503.01864/) (2025)
- [The Quest for Efficient Reasoning: A Data-Centric Benchmark to CoT Distillation](https://mlsanity.com/paper/2505.18759/) (2025)
- [Who's Your Judge? On the Detectability of LLM-Generated Judgments](https://mlsanity.com/paper/2509.25154/) (2025)
- [LLM Safety Alignment is Divergence Estimation in Disguise](https://mlsanity.com/paper/2502.00657/) (2025)
- [GuiLoMo: Allocating Expert Number and Rank for LoRA-MoE via Bilevel Optimization with GuidedSelection Vectors](https://mlsanity.com/paper/2506.14646/) (2025)
- [Diagnosis Before Recovery: Turning Agent Failures into Selective Self-Correction](https://mlsanity.com/paper/2608.11772/) (2026)
- [Assessing the Impact of Conspiracy Theories Using Large Language Models](https://mlsanity.com/paper/2412.07019/) (2024)
- [From Fragments to Facts: A Curriculum-Driven DPO Approach for Generating Hindi News Veracity Explanations](https://mlsanity.com/paper/2507.05179/) (2025)
- [How to Train a Real-World Silicon Concierge? Internalizing Complex Business Workflow to Only OneModel](https://mlsanity.com/paper/2608.20350/) (2026)
- [CURA: Clinical Uncertainty Risk Alignment for Language Model-Based Risk Prediction](https://mlsanity.com/paper/2604.14651/) (2026)
- [ARC: Fair Relative Advantage Comparison in Open-Ended Real-World Interaction](https://mlsanity.com/paper/2608.13622/) (2026)
- [Ask, Condition or Abstain: Reinforcement Learning for Missing-Premise Reasoning](https://mlsanity.com/paper/2608.16554/) (2026)

## Abstract (arXiv)

Reinforcement Learning with Human Feedback (RLHF) is the key to the success of large language models (LLMs) in recent years. In this work, we first introduce the concepts of knowledge breadth and knowledge depth, which measure the comprehensiveness and depth of an LLM or knowledge source respectively. We reveal that the imbalance in the number of prompts and responses can lead to a potential disparity in breadth and depth learning within alignment tuning datasets by showing that even a simple uniform method for balancing the number of instructions and responses can lead to significant improvements. Building on this, we further propose Balanced Preference Optimization (BPO), designed to dynamically augment the knowledge depth of each sample. BPO is motivated by the observation that the usefulness of knowledge varies across samples, necessitating tailored learning of knowledge depth. To achieve this, we introduce gradient-based clustering, estimating the knowledge informativeness and usefulness of each augmented sample based on the model's optimization direction. Our experimental results across various benchmarks demonstrate that BPO outperforms other baseline methods in alignment tuning while maintaining training efficiency. Furthermore, we conduct a detailed analysis of each component of BPO, providing guidelines for future research in preference data optimization.