Fine-Tuning

Follow this topic Ask about this topic

Fine-tuning starts from a pretrained model and uses task-specific data or feedback to adapt its predictions, representations, or behavior. Approaches range from updating all weights to parameter-efficient methods such as LoRA and adapters, selective layer or neuron updates, and supervised, preference-based, or reinforcement-learning objectives. The techniques are used across language, vision, speech, robotics, and scientific applications, with evaluation spanning task accuracy, domain transfer, safety, and deployment costs. A central challenge is gaining new capabilities without losing useful prior knowledge: outcomes depend on the model, data, and training choices, and improvements on one benchmark may not transfer to other settings.

Latest papers

250 papers23 papers in the last four weeks, down 30% on the four weeks before. 0.1% of all new papers.

Rescuing Your Pruning: Rethinking Visual Token Pruning and Post-Training Recovery in MLLMs
Oct 24, 2026Peiran Yin, Hao Wu, Zihao Yang +3Visual Token Pruning Efficient VLM Inference
EIT-NLP Lab, Eastern Institute of Technology, Ningbo · Shanghai Jiao Tong University · Uppsala University
Propensities Please: Exposing Alignment-Faking LLMs using Synthetic Document Finetuning
Oct 23, 2026Ardy Haroen, Florian Lorkowski, Samuel Nellessen +1Language Model Safety Evaluation LLM Auditing
LASR Labs · Apollo Research
DRIVE: Fine-tuning via Data Contribution- and Diversity-aware Weighting with Prior Regularization
Oct 23, 2026Qing Liu, Xinrui Chen, Yi Du +2Fine-Tuning Instance Reweighting
Hangzhou Institute for Advanced Study, University of Chinese Academy of Sciences, Hangzhou, China · Computer Network Information Center, Chinese Academy of Sciences, Beijing, China · School of Mathematics, Tianjin University, Tianjin, China
Latent space configuration for neural network finetuning, fast ranking, and embedding dimensionality reduction
Oct 21, 2026Nikita Gabdullin, Anton RaskovalovRepresentation Learning Fine-Tuning
Joint Stock "Research and production company "Kryptonite"
Is it RL or Data? A Controlled Comparison of GRPO and Rejection-Sampling Fine-Tuning
Oct 19, 2026Lakshya ChaudhryFine-Tuning RL Fine-Tuning
Purdue University
Confine to the Task, Anchor the Rest: Concept-Guided Retention Regularizer for Parameter-Efficient Fine-Tuning
Oct 13, 2026Ibna Kowsar, Ivan Medri, Fanyu Meng +1Fine-Tuning Parameter-Efficient Fine-Tuning
Department of Computer Science, University of California, Davis · University of Virginia
Select or Rewrite: Oracle-Guided Fine-Tuning for Generative ASR Post-Editing from N-Best Hypotheses
Oct 10, 2026Shiran Liu, Gaohui Xu, Sergio Burdisso +3Language Model Post-Training Fine-Tuning
Idiap Research Institute, Martigny, Switzerland · KTH Royal Institute of Technology, Stockholm, Sweden · Brno University of Technology, Brno, Czech Republic
An Investigation of Model Coherence: Narrow Finetunes Contradict Themselves Under Resampling
Oct 8, 2026Robert Graham, Yariv Barsheshat, Phil Blandfort +1Reasoning Consistency in Language Models Self-Consistency in Language Models
Independent · Predictably Weird · Amazon
MAP4CS: A Multi-dimensional Data Pruning Framework for Efficient Code Retriever Fine-tuning
Oct 8, 2026Yuxuan Chen, Mingwei Liu, Guangsheng Ou +4Code Search Training Data Selection
School of Software Engineering, Sun Yat-sen University, China
SP-DocReader: Difference-Aware Self-Play for Precise Document OCR
Oct 8, 2026Wenjie Liao, Xiaohui Song, Liangjie Zhao +1Optical Character Recognition Vision-Language Models
Guangdong OPPO Mobile Telecommunications Corp.,Ltd · Adelaide University
Stability-Plasticity Balance via Singular-Vector Selection in LLM Continual Learning
Oct 8, 2026Lingxiang Wang, Hainan Zhang, Liang Pang +2Continual Learning for LLMs Parameter-Efficient Fine-Tuning
School of Artificial Intelligence, Beihang University · Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing · Institute of Computing Technology, Chinese Academy of Sciences
Comprehensive Evaluation and Fine-Tuning of Foundational Cell Nuclei Segmentation Models in Renal Pathology
Oct 6, 2026Ruijie Wu, Junlin Guo, Ruining Deng +4Medical Imaging Foundation Models Fine-Tuning
Department of Electrical and Computer Engineering, Vanderbilt University, Nashville, TN 37235, USA · Department of Radiology, Weill Cornell Medicine, New York, NY 10021, USA · Department of Biostatistics, Vanderbilt University Medical Center, Nashville, TN 37235, USA +2
HeuFouFT: Task-Guided Metaheuristic Coordinate Search for Fourier Fine-Tuning
Oct 5, 2026Ruiheng Wang, Yubo Hou, Yakun Zhu +3Fine-Tuning LLM Fine-Tuning
Beihang University, Beijing, China · Shanghai Jiao Tong University, Shanghai, China · VinUniversity, Hanoi, Vietnam
Pre-training interventions, ex post facto: Grafting model beliefs across checkpoints
Sep 30, 2026Peter Nutter, Dani Roytburg, Clément Dumas +2Fine-Tuning LLM Alignment
ETH Zurich · MATS · Carnegie Mellon University +2
Bayesian Fine-tuning Yields Language Models that are as Bayesian as their Beliefs Allow
Sep 30, 2026Polina Tsvilodub, Andreas Waldis, Linlu Qiu +2Fine-Tuning LLM Fine-Tuning
University of Tübingen · Massachusetts Institute of Technology · New York University
Fenchel Tilting: Weighted Correction for Efficient Finetuning of Generative Models
Sep 30, 2026Maksim Bobrin, Maksim Zhdanov, Dmitry DylovDiffusion Model Alignment Constrained Generative Modeling
AXXX · Applied AI Institute · MBZUAI
Alignment via Training Against Probes Without Losing Monitorability
Sep 29, 2026Lena Libon, Alexander Panfilov, Ben Rank +3Fine-Tuning LLM Safety Alignment
ETH Zurich · ELLIS Institute Tübingen, MPI for Intelligent Systems, Tübingen AI Center
On-Policy Parameter Update Direction Underlies Generalization in LLM Post-Training
Sep 29, 2026Shufan Shen, Zhongni Hou, Junshu Sun +5LLM Post-Training Fine-Tuning
State Key Lab. of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences · University of Chinese Academy Sciences · Meituan
Narrow Multimodal Fine-Tuning Can Induce Emergent Misalignment
Sep 28, 2026Shunchang Liu, Lukas Fluri, Xin Chen +1Vision-Language Models Fine-Tuning
ETH Zürich · EPFL · Aalto University +1
Layer-Informed Fine-Tuning via Three-Stage Functional Segmentation of LLMs
Sep 28, 2026Junning Shao, Siwei Wang, Zhixuan FangLLM Fine-Tuning Fine-Tuning
Tsinghua University, Beijing, China · Microsoft Research Asia, Beijing, China · Shanghai Qi Zhi Institute, Shanghai, China
Statistical Benefits of Fine-Tuning from Pretrained Initialization in Diagonal Linear Networks
Sep 28, 2026Alexandre Declèves, Etienne Boursier, Nicolas FlammarionFine-Tuning Deep Linear Networks
TML, EPFL, Lausanne, Switzerland · INRIA, LMO, Université Paris-Saclay, Orsay, France
Two-Timescale Fine-tuning Provably Learns New Features for Two-Layer ReLU Networks
Sep 28, 2026Etienne Boursier, Nicolas FlammarionRepresentation Learning Fine-Tuning
INRIA, LMO, Universit´e Paris-Saclay, Orsay, France · EPFL, Lausanne, Switzerland
MaskCoFT: Masked Co-Adaptive Fine-Tuning for Memory-Efficient MoE Inference
Sep 28, 2026Junfeng Wu, Zehao Fan, Hadjer Benmeziane +3Expert Offloading Mixture-of-Experts Language Models
Department of Industrial and Systems Engineering, Rensselaer Polytechnic Institute · Department of Electrical, Computer, and Systems Engineering, Rensselaer Polytechnic Institute · IBM Research
Resource-Aware Parameter-Efficient Model Adaptation for Onboard High-Dimensional Data
Sep 27, 2026Qiyang Zhang, Xinhao Li, Lei Shi +4Fine-Tuning Low-Rank Adaptation
Beijing University of Posts and Telecommunications · Wuhan University · Communication University of China +1
A Light Bilevel Refinement Aligns Self-Supervised Representations for Stronger Task-Specific Learning
Sep 27, 2026Gustav Wagner Zakarias, Zheng-Hua TanFine-Tuning Bilevel Optimization
Aalborg University · Pioneer Centre for Artificial Intelligence, Denmark
Reuse or Relearn? A Spectral View of Earth Observation Foundation Models
Sep 26, 2026Mehmet Ozgur Turkoglu, Valerio Marsocci, Dominik J. Mühlematter +3Fine-Tuning Geospatial Foundation Models
Agroscope, EO of Agroecosystems · ESA, Φ-lab · ETH Zurich
SoFT: Soft Targets for Generalizable LLM Fine-Tuning
Sep 26, 2026Huihao Jing, Wenbin Hu, Shaojin Chen +7LLM Fine-Tuning Fine-Tuning
HKUST, HKUST-GZ
What Should We Freeze? Guarded Freezing: Connectivity Shapes the Fine-Tuning of Pretrained Models
Sep 26, 2026Leonel AguilarFine-Tuning Selective Fine-Tuning
Chair of Cognitive Science, ETH Zürich
Can One Adapted Model Do It All? Fine-Tuning Strategy Selection for Customer Support LLMs
Sep 23, 2026Md Tahmid Rahman Laskar, Xue-Yong Fu, Shashi Bhushan TNFine-Tuning Customer Support Automation
Dialpad Inc.
MIS-Bench: Benchmarking Multimodal LLMs for Psychotherapeutic Interpersonal Skills Assessment
Sep 19, 2026Yuhan Lu, Yi Yao, Hua Shen +2Clinical Language Model Evaluation Multimodal Model Evaluation
NYU Shanghai, New York University

Oct 24, 2026J/K move · Enter open · S save

Rescuing Your Pruning: Rethinking Visual Token Pruning and Post-Training Recovery in MLLMs

Peiran Yin, Hao Wu, Zihao Yang, Haozhe Hu+2

EIT-NLP Lab, Eastern Institute of Technology, Ningbo · Shanghai Jiao Tong University · Uppsala University

Visual Token Pruning Efficient VLM Inference Visual Token Compression VLM Post-Training Fine-Tuning

Open paper AskPDF arXiv

Visual token pruning substantially reduces the inference cost of multimodal large language models (MLLMs), yet existing methods are mostly training-free and plug-and-play. They treat the resulting performance drop as an inevitable cost of compression and pursue increasingly complex pruning rules. We rethink this view through two questions: how much of the loss can post-training rescue, and, building on this recovery, how can joint compression push pruning to its limit? We curate a 282K-sample post-training mixture covering general VQA, reasoning, OCR, and grounding, and post-train 36 configurations that drop or merge tokens under a unified protocol. Post-training rescues most of the loss, even for OCR and grounding, and its adaptation gain in closing the gap between the pruned and the dense model grows as the budget tightens. For joint compression, we find that the vision encoder, not the LLM, dominates prefill time in native-resolution MLLMs, so where tokens are removed matters more than how many: keeping 40% of tokens at the encoder runs faster than keeping 10% at the projector. Our best configuration prunes the encoder lightly and the projector aggressively, leaves the LLM intact, and keeps general understanding and reasoning at the dense level. It accelerates the LLM stage by 3.49x but the end-to-end time-to-first-token by only 1.54x, showing that LLM-stage speedups overstate end-to-end gains. We hope these findings offer insights for future research on joint visual token compression. Our data, protocol, and code will be released.