An Empirical Study of SFT-DPO Interaction and Parameterization in Small Language Models

• URL: http://arxiv.org/abs/2603.20100v1
• Date: Fri, 20 Mar 2026 16:27:03 GMT
• Title: An Empirical Study of SFT-DPO Interaction and Parameterization in Small Language Models
• Abstract summary: We compare SFT-only, DPO-only, and staged SFT-to-DPO training alongside full fine-tuning (FFT) versus LoRA on a GPT-2-scale decoder.<n>DPO yields small, task-dependent gains over strong SFT and can match competitive SFT accuracy without a warm start.
• Score: 1.7559669963238091
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Direct Preference Optimization (DPO) is widely used after supervised fine-tuning (SFT) to align language models, yet empirical behavior under small backbones and modest data is under-specified. We systematically compare SFT-only, DPO-only, and staged SFT-to-DPO training alongside full fine-tuning (FFT) versus LoRA on a GPT-2-scale decoder, evaluating paraphrase detection and Shakespearean sonnet continuation. DPO yields small, task-dependent gains over strong SFT and can match competitive SFT accuracy without a warm start when the preference construction closely parallels the supervised objective. In contrast, parameterization dominates: FFT consistently outperforms LoRA at matched training depth, and LoRA does not reduce wall-clock time on our hardware. These findings indicate that, in this small-scale regime, supervised full-parameter adaptation remains the primary performance lever, while preference optimization and low-rank adaptation provide limited marginal returns.

Related papers

• Trust-Region Adaptive Policy Optimization [82.09255251747818]
  Post-training methods play an important role in improving large language models' (LLMs) complex reasoning abilities.<n>We introduce TRAPO, a framework that interleavesSupervised Fine-Tuning (SFT) and Reinforcement Learning (RL) within each training instance.<n>Experiments on five mathematical reasoning benchmarks show that TRAPO consistently surpasses standard SFT, RL, and SFT-then-RL pipelines.
  arXiv  Detail & Related papers  (2025-12-19T14:37:07Z)
• Self-Rewarding PPO: Aligning Large Language Models with Demonstrations Only [70.43369087819332]
  Supervised fine-tuning (SFT) has emerged as a crucial method for aligning large language models with human-annotated demonstrations.<n>We propose Self-Rewarding PPO, a novel fine-tuning method that leverages on-policy techniques to enhance generalization performance.
  arXiv  Detail & Related papers  (2025-10-24T02:02:13Z)
• Implicit Reward as the Bridge: A Unified View of SFT and DPO Connections [65.36449542323277]
  We present a unified theoretical framework bridgingSupervised Fine-Tuning (SFT) and preference learning in Large Language Model (LLM) post-training.<n>We propose a simple yet effective learning rate reduction approach that yields significant performance improvements.
  arXiv  Detail & Related papers  (2025-06-15T05:42:29Z)
• Discriminative Finetuning of Generative Large Language Models without Reward Models and Human Preference Data [73.04828796123581]
  Supervised fine-tuning (SFT) has become a crucial step for aligning pretrained large language models (LLMs)<n>We introduce Discriminative Fine-Tuning (DFT), an improved variant of SFT, which mitigates the burden of collecting human-labeled preference data.<n>Our contributions include: (i) a discriminative probabilistic framework for fine-tuning LLMs by explicitly modeling the discriminative likelihood of an answer among all possible outputs given an input; (ii) efficient algorithms to optimize this discriminative likelihood; and (iii) extensive experiments demonstrating DFT's effectiveness
  arXiv  Detail & Related papers  (2025-02-25T22:38:55Z)
• Bridging SFT and DPO for Diffusion Model Alignment with Self-Sampling Preference Optimization [67.8738082040299]
  Self-Sampling Preference Optimization (SSPO) is a new alignment method for post-training reinforcement learning.<n>SSPO eliminates the need for paired data and reward models while retaining the training stability of SFT.<n>SSPO surpasses all previous approaches on the text-to-image benchmarks and demonstrates outstanding performance on the text-to-video benchmarks.
  arXiv  Detail & Related papers  (2024-10-07T17:56:53Z)
• ASFT: Aligned Supervised Fine-Tuning through Absolute Likelihood [14.512464277772194]
  Aligned Supervised Fine-Tuning (ASFT) is an effective approach that better aligns Large Language Models with pair-wise datasets. ASFT mitigates the issue where the DPO loss function decreases the probability of generating human-dispreferred data. Extensive experiments demonstrate that ASFT is an effective alignment approach, consistently outperforming existing methods.
  arXiv  Detail & Related papers  (2024-09-14T11:39:13Z)
• Minor SFT loss for LLM fine-tune to increase performance and reduce model deviation [9.506166330956082]
  We propose a training metric for SFT to measure the discrepancy between the optimized model and the original model, and a loss function MinorSFT that can increase the training effectiveness. In this article with insight from DPO and MinorDPO, we propose a training metric for SFT to measure the discrepancy between the optimized model and the original model, and a loss function MinorSFT that can increase the training effectiveness.
  arXiv  Detail & Related papers  (2024-08-20T08:32:44Z)
• Intuitive Fine-Tuning: Towards Simplifying Alignment into a Single Process [19.986235452236272]
  Supervised Fine-Tuning (SFT) and Preference Optimization (PO) are key processes for aligning Language Models (LMs) with human preferences post pre-training.<n>We introduce Intuitive Fine-Tuning (IFT) to integrate SFT and PO into a single process.<n>IFT performs comparably or even superiorly to SFT and some typical PO methods across several tasks.
  arXiv  Detail & Related papers  (2024-05-20T08:23:28Z)
• The Gaps between Pre-train and Downstream Settings in Bias Evaluation and Debiasing [74.7319697510621]
  In-Context Learning (ICL) induces smaller changes to PLMs compared to FT-based debiasing methods. ICL-based debiasing methods show a higher correlation between intrinsic and extrinsic bias scores compared to FT-based methods.
  arXiv  Detail & Related papers  (2024-01-16T17:15:08Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.