# LP-SFT: Keeping Plausible Alternatives Alive in Supervised Fine-Tuning

arXiv: https://arxiv.org/abs/2607.04733 · HTML: https://mlsanity.com/paper/2607.04733/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 24 references that are in the map.
- First posted on arXiv 6 Jul 2026.

## Builds on (most cited first, 20 of 20)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [Self-Instruct: Aligning Language Models with Self-Generated Instructions](https://mlsanity.com/paper/2212.10560/) (2022)
- [Is Your Code Generated by ChatGPT Really Correct? Rigorous Evaluation of Large Language Models for Code Generation](https://mlsanity.com/paper/2305.01210/) (2023)
- [On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes](https://mlsanity.com/paper/2306.13649/) (2023)
- [An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-tuning](https://mlsanity.com/paper/2308.08747/) (2023)
- [Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning](https://mlsanity.com/paper/2506.01939/) (2025)
- [Neural Text Generation with Unlikelihood Training](https://mlsanity.com/paper/1908.04319/) (2019)
- [UltraFeedback: Boosting Language Models with Scaled AI Feedback](https://mlsanity.com/paper/2310.01377/) (2023)
- [Decoupled Knowledge Distillation](https://mlsanity.com/paper/2203.08679/) (2022)
- [Magicoder: Empowering Code Generation with OSS-Instruct](https://mlsanity.com/paper/2312.02120/) (2023)
- [On the Generalization of SFT: A Reinforcement Learning Perspective with Reward Rectification](https://mlsanity.com/paper/2508.05629/) (2025)
- [Preserving Diversity in Supervised Fine-Tuning of Large Language Models](https://mlsanity.com/paper/2408.16673/) (2024)
- [Entropy-Adaptive Fine-Tuning: Resolving Confident Conflicts to Mitigate Forgetting](https://mlsanity.com/paper/2601.02151/) (2026)
- [HE-SNR: Uncovering Latent Logic via Entropy for Guiding Mid-Training on SWE-bench](https://mlsanity.com/paper/2601.20255/) (2026)

## Built on by (most cited first, 1 of 1)

- [Beyond Distribution Matching: Selective Margin Preservation for Supervised Fine-Tuning](https://mlsanity.com/paper/2610.19685/) (2026)

## Abstract (arXiv)

Supervised fine-tuning (SFT) is widely used to adapt pretrained language models to downstream domains, but can over-specialize the model and degrade its pre-existing capabilities. Standard cross-entropy drives probability mass onto the observed target token and can make plausible alternatives that the pretrained model itself endorsed vanish. Using Shannon and Renyi entropies, we show that pretrained next-token distributions exhibit a regular multimodal structure, with entropy peaks corresponding to different numbers of plausible alternatives. Motivated by this observation, we propose LP-SFT, a Local-Preserving Supervised Fine-Tuning objective with two key design principles: it removes the supervised token from a local top-K candidate set to avoid conflict with cross-entropy, and applies a locally normalized KL loss to preserve relative preferences among the remaining non-label alternatives. Across mixed-domain and domain-specific fine-tuning experiments, LP-SFT consistently outperforms standard SFT and recent baselines in aggregate performance while keeping base-endorsed alternatives from vanishing, achieving a favorable balance between single-sample accuracy and finite-budget solution accessibility, as measured by pass@1 and pass@k, respectively, with only a modest increase in training cost.