Thinking Preference Optimization

TL;DR

Thinking Preference Optimization (ThinkPO) is a post-training method that improves long chain-of-thought reasoning in language models by encouraging longer reasoning outputs through preference optimization, without needing new data.

Contribution

The paper introduces ThinkPO, a simple method that enhances reasoning performance of SFT models by leveraging existing short and long CoT responses for preference optimization.

Findings

Increases math reasoning accuracy by 8.6%.

Boosts output length by 25.9%.

Improves the performance of distilled SFT models on benchmarks.

Abstract

Supervised Fine-Tuning (SFT) has been a go-to and effective method for enhancing long chain-of-thought (CoT) reasoning in relatively small LLMs by fine-tuning them with long CoT responses from larger LLMs. To continually improve reasoning abilities, we can either collect new high-quality long CoT reasoning SFT data or repeatedly train on existing SFT datasets. However, acquiring new long CoT SFT data is costly and limited, while repeated training often results in a performance plateau or decline. To further boost the performance with the SFT data, we propose Thinking Preference Optimization (ThinkPO), a simple yet effective post-SFT method that enhances long CoT reasoning without requiring new long CoT responses. Instead, ThinkPO utilizes readily available or easily obtainable short CoT reasoning responses as rejected answers and long CoT responses as chosen answers for the same question. It then applies direct preference optimization to encourage the model to favor longer reasoning outputs. Experiments show that ThinkPO further improves the reasoning performance of SFT-ed models, e.g. it increases math reasoning accuracy of SFT-ed models by 8.6% and output length by 25.9%. Notably, ThinkPO is capable of continually boosting the performance of the publicly distilled SFT model, e.g., increasing the official DeepSeek-R1-Distill-Qwen-7B's performance on MATH500 from 87.4% to 91.2%.