Cover Image

When Inverse Data Outperforms: Exploring the Pitfalls of Mixed Data in Multi-Stage Fine-Tuning

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Deng, Mengyi, Li, Xin, Zhu, Tingyu, Yang, Zhicheng, Guo, Zhijiang, Wang, Wei
Format:	Preprint
Published:	2025
Subjects:	Machine Learning Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Efficient RLVR Training via Weighted Mutual Information Data Selection
by: Zhou, Xinyu, et al.
Published: (2026)

DGPO: Beyond Pairwise Preferences with Directional Consistent Groupwise Optimization
by: Deng, Mengyi, et al.
Published: (2026)

EnvFactory: Scaling Tool-Use Agents via Executable Environments Synthesis and Robust RL
by: Xu, Minrui, et al.
Published: (2026)

What Matters in LLM-generated Data: Diversity and Its Effect on Model Fine-Tuning
by: Zhu, Yuchang, et al.
Published: (2025)

Learning Like Humans: Resource-Efficient Federated Fine-Tuning through Cognitive Developmental Stages
by: Wu, Yebo, et al.
Published: (2025)

Supervised Fine-Tuning as Inverse Reinforcement Learning
by: Sun, Hao
Published: (2024)

SRFT: A Single-Stage Method with Supervised and Reinforcement Fine-Tuning for Reasoning
by: Fu, Yuqian, et al.
Published: (2025)

Mind the Gap: Data Rewriting for Stable Off-Policy Supervised Fine-Tuning
by: Zhao, Shiwan, et al.
Published: (2025)

ClusterUCB: Efficient Gradient-Based Data Selection for Targeted Fine-Tuning of LLMs
by: Wang, Zige, et al.
Published: (2025)

Why Softmax Attention Outperforms Linear Attention
by: Deng, Yichuan, et al.
Published: (2023)

Task-Specific Efficiency Analysis: When Small Language Models Outperform Large Language Models
by: Cao, Jinghan, et al.
Published: (2026)

Selection of LLM Fine-Tuning Data based on Orthogonal Rules
by: Li, Xiaomin, et al.
Published: (2024)

Anchored Supervised Fine-Tuning
by: Zhu, He, et al.
Published: (2025)

SFT-then-RL Outperforms Mixed-Policy Methods for LLM Reasoning
by: Limozin, Alexis, et al.
Published: (2026)

From Instance Selection to Fixed-Pool Data Recipe Search for Supervised Fine-Tuning
by: Wu, Haodong, et al.
Published: (2026)

Mix, Don't Tune: Bilingual Pre-Training Outperforms Hyperparameter Search in Data-Constrained Settings
by: Jeha, Paul, et al.
Published: (2026)

Cream of the Crop: Harvesting Rich, Scalable and Transferable Multi-Modal Data for Instruction Fine-Tuning
by: Lyu, Mengyao, et al.
Published: (2025)

Data Diversity Matters for Robust Instruction Tuning
by: Bukharin, Alexander, et al.
Published: (2023)

From Parameters to Data: A Task-Parameter-Guided Fine-Tuning Pipeline for Efficient LLM Alignment
by: Chen, Hao, et al.
Published: (2026)

Can LLMs $\textit{understand}$ Math? -- Exploring the Pitfalls in Mathematical Reasoning
by: Roy, Tiasa Singha, et al.
Published: (2025)

From Unstructured Data to In-Context Learning: Exploring What Tasks Can Be Learned and When
by: Wibisono, Kevin Christian, et al.
Published: (2024)

Exploring Parameter-Efficient Fine-Tuning Techniques for Code Generation with Large Language Models
by: Weyssow, Martin, et al.
Published: (2023)

Fine-Tuning vs. RAG for Multi-Hop Question Answering with Novel Knowledge
by: Yang, Zhuoyi, et al.
Published: (2026)

FisherSFT: Data-Efficient Supervised Fine-Tuning of Language Models Using Information Gain
by: Deb, Rohan, et al.
Published: (2025)

Synthetic Data Generation in Low-Resource Settings via Fine-Tuning of Large Language Models
by: Kaddour, Jean, et al.
Published: (2023)

Fine-Tuning LLMs for Report Summarization: Analysis on Supervised and Unsupervised Data
by: Rallapalli, Swati, et al.
Published: (2025)

Filter-then-Weight: Online Data Selection and Reweighting for LLM Fine-Tuning
by: Wang, Fangxin, et al.
Published: (2026)

Diversity as a Reward: Fine-Tuning LLMs on a Mixture of Domain-Undetermined Data
by: Ling, Zhenqing, et al.
Published: (2025)

Mix Data or Merge Models? Optimizing for Diverse Multi-Task Learning
by: Aakanksha, et al.
Published: (2024)

GiFT: Gibbs Fine-Tuning for Code Generation
by: Li, Haochen, et al.
Published: (2025)

Neural Networks for Learnable and Scalable Influence Estimation of Instruction Fine-Tuning Data
by: Agarwal, Ishika, et al.
Published: (2025)

BEFT: Bias-Efficient Fine-Tuning of Language Models in Low-Data Regimes
by: Huang, Baichuan, et al.
Published: (2025)

Exploring Fine-Tuning of Large Audio Language Models for Spoken Language Understanding under Limited Speech Data
by: Choi, Youngwon, et al.
Published: (2025)

MoRA: High-Rank Updating for Parameter-Efficient Fine-Tuning
by: Jiang, Ting, et al.
Published: (2024)

PoliTune: Analyzing the Impact of Data Selection and Fine-Tuning on Economic and Political Biases in Large Language Models
by: Agiza, Ahmed, et al.
Published: (2024)

MMICT: Boosting Multi-Modal Fine-Tuning with In-Context Examples
by: Chen, Tao, et al.
Published: (2023)

Scaling Laws for Forgetting When Fine-Tuning Large Language Models
by: Kalajdzievski, Damjan
Published: (2024)

Federated Data-Efficient Instruction Tuning for Large Language Models
by: Qin, Zhen, et al.
Published: (2024)

Prompt Tuning for Natural Language to SQL with Embedding Fine-Tuning and RAG
by: Jang, Jisoo, et al.
Published: (2025)

When Does Predictive Inverse Dynamics Outperform Behavior Cloning?
by: Schäfer, Lukas, et al.
Published: (2026)