# Logits-Based Finetuning

arXiv: https://arxiv.org/abs/2505.24461 · HTML: https://mlsanity.com/paper/2505.24461/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 30 May 2025.

## Built on by (most cited first, 3 of 3)

- [Uni-OPD: Unifying On-Policy Distillation with a Dual-Perspective Recipe](https://mlsanity.com/paper/2605.03677/) (2026)
- [VisionThink: Smart and Efficient Vision Language Model via Reinforcement Learning](https://mlsanity.com/paper/2507.13348/) (2025)
- [Uni-DPO: A Unified Paradigm for Dynamic Preference Optimization of LLMs](https://mlsanity.com/paper/2506.10054/) (2025)

## Abstract (arXiv)

In recent years, developing compact and efficient large language models (LLMs) has emerged as a thriving area of research. Traditional Supervised Fine-Tuning (SFT), which relies on singular ground truth labels, often fails to capture token-level dependencies and linguistic diversity. To address these limitations, we propose a logits-based fine-tuning framework that integrates the strengths of supervised learning and knowledge distillation. Our approach constructs enriched training targets by combining teacher logits with ground truth labels, preserving both correctness and linguistic diversity. This ensures more reliable and effective training. We constructed a large-scale 1.2M logits dataset and trained a series of science-focused models. Experimental results demonstrate that our method achieves significant improvements, with accuracy gains of 18% on Mawps and 22.7% on TabMWP. Across nine widely used mathematical benchmarks, our method consistently outperforms prior SFT models, achieving an average improvement of 7.28%. Codes are available at https://github.com/dvlab-research/Logits-Based-Finetuning.