# Reclaiming Low-Quality Instruction Data with Rule-Guided LLM Reconstruction

arXiv: https://arxiv.org/abs/2510.23160 · HTML: https://mlsanity.com/paper/2510.23160/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 27 Oct 2025.

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

- [Replay What Matters: Off-Policy Replay for Efficient LLM Reinforcement Unlearning](https://mlsanity.com/paper/2606.15333/) (2026)

## Abstract (arXiv)

Data selection for supervised fine-tuning (SFT) often discards low-scoring instruction examples that may retain useful supervision. We propose a rule-guided LLM reconstruction framework to reclaim this information. The framework identifies low-scoring samples through LLM-based rating and score correction, selects semantically representative sources, and reconstructs instruction-response pairs through strategy-guided generation, criteria-based validation, and failure-guided repair. From 123,786 low-scoring samples, it constructs 54,888 training examples. Across three 7B/8B base LLMs, the reconstructed data achieves higher average downstream performance than the low-scoring source set and all evaluated data-selection baselines. Despite using substantially fewer examples than the full source dataset (approximately 300K), it achieves higher reported average scores than full-data training across all three base models. Stepwise ablations identify rule-guided reconstruction as the largest source of observed improvements over direct fusion, while a manual audit reveals remaining limitations in source integration and answer faithfulness. These findings support reconstruction as a complementary approach to selection-based SFT data curation.