paper-with-me

Papers

Self-Training with Direct Preference Optimization Improves Chain-of-Thought Reasoning

2024-07-25 · Tianduo Wang, Shichen Li, Wei Lu

Effective training of language models (LMs) for mathematical reasoning tasks demands high-quality supervised fine-tuning data. Besides obtaining annotations from human experts, a common alternative is sampling from larger and more powerful LMs. However, this knowledge distillation approach can be costly and unstable, particularly when relying on closed-source, proprietary LMs like GPT-4, whose behaviors are often unpredictable. In this work, we demonstrate that the reasoning abilities of small-scale LMs can be enhanced through self-training, a process where models learn from their own outputs. We also show that the conventional self-training can be further augmented by a preference learning algorithm called Direct Preference Optimization (DPO). By integrating DPO into self-training, we leverage preference data to guide LMs towards more accurate and diverse chain-of-thought reasoning. We evaluate our method across various mathematical reasoning tasks using different base models. Our experiments show that this approach not only improves LMs' reasoning performance but also offers a more cost-effective and scalable solution compared to relying on large proprietary LMs.

📄 PDF Abstract BibTeX arXiv:2407.18248

Code (1)

tianduowang/dpo-st 공식 구현 pytorch

Tasks

Knowledge DistillationMathematical Reasoning

Methods 이 논문이 사용한 방법론

Attention 설명 없음
DPO 설명 없음
Adam 설명 없음
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…

Similar Papers 제목 키워드 기반

SGPO: Self-Generated Preference Optimization based on Self-Improver

2025-07-27 · Hyeonji Lee, Daejin Jo, Seohwan Yun, Sungwoong Kim arxiv

Large language models (LLMs), despite their extensive pretraining on diverse datasets, require effective alignment to human preferences for practical and reliable deployment. Conventional alignment methods typically empl…

SEE-DPO: Self Entropy Enhanced Direct Preference Optimization

2024-11-06 · Shivanshu Shekhar, Shreyas Singh, Tong Zhang

Direct Preference Optimization (DPO) has been successfully used to align large language models (LLMs) according to human preferences, and more recently it has also been applied to improving the quality of text-to-image d…

DiversityImage GenerationSpecificity

Revisiting Self-Play Preference Optimization: On the Role of Prompt Difficulty

2025-10-07 · Yao Xiao, Jung-jae Kim, Roy Ka-wei Lee, Lidong Bing arxiv

Self-play preference optimization has emerged as a prominent paradigm for aligning large language models (LLMs). It typically involves a language model to generate on-policy responses for prompts and a reward model (RM) …

Self-Evolutionary Large Language Models through Uncertainty-Enhanced Preference Optimization

2024-09-17 · Jianing Wang, Yang Zhou, Xiaocheng Zhang, Mengjiao Bao 외

Iterative preference optimization has recently become one of the de-facto training paradigms for large language models (LLMs), but the performance is still underwhelming due to too much noisy preference data yielded in t…

Token-weighted Direct Preference Optimization with Attention

2026-05-21 · Chengyu Huang, Zhuohang Li, Sheng-Yen Chou, Claire Cardie arxiv

Direct Preference Optimization (DPO) aligns Large Language Models with human preferences without the need for a separate reward model. However, DPO treats all tokens in responses equally, neglecting the differing importa…