mlsanityData to 14 Oct 2026

Propulsion: Steering LLM with Tiny Fine-Tuning

arXiv 2409.10927 ↗First posted 17 Sep 2024cs.CLcited by 14

Not on the map. Why?

Abstract (arXiv)

The rapid advancements in Large Language Models (LLMs) have revolutionized natural language processing (NLP) and related fields. However, fine-tuning these models for specific tasks remains computationally expensive and risks degrading pre-learned features. To address these challenges, we propose Propulsion, a novel parameter efficient fine-tuning (PEFT) method designed to optimize task-specific performance while drastically reducing computational overhead. Inspired by the concept of controlled adjustments in physical motion, Propulsion selectively re-scales specific dimensions of a pre-trained model, guiding output predictions toward task objectives without modifying the model's parameters. By introducing lightweight, trainable Propulsion parameters at the pre-trained layer, we minimize the number of parameters updated during fine-tuning, preventing overfitting or overwriting of existing knowledge. Our theoretical analysis, supported by Neural Tangent Kernel (NTK) theory, shows that Propulsion approximates the performance of full fine-tuning with far fewer trainable parameters. Empirically, Propulsion reduces the parameter count from 355.3 million to just 0.086 million, achieving over a 10x reduction compared to standard approaches like LoRA while maintaining competitive performance across benchmarks.

Key facts. Cited by 14 papers in the map (data through 14 Oct 2026). Builds on 61 of its 90 references that are in the map. First posted on arXiv 17 Sep 2024.

Builds on

61 of 90 references.

References · 61

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020Language Models are Few-Shot Learnerscited by 32,419
  3. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  4. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  5. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  6. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  7. 2019BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehensioncited by 5,893
  8. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,899
  9. 2018Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challengecited by 4,886
  10. 2019Parameter-Efficient Transfer Learning for NLPcited by 4,236
  11. 2016SQuAD: 100,000+ Questions for Machine Comprehension of Textcited by 4,148
  12. 2019HellaSwag: Can a Machine Really Finish Your Sentence?cited by 4,100
38 more
  1. 2021Prefix-Tuning: Optimizing Continuous Prompts for Generationcited by 3,915
  2. 2021The Power of Scale for Parameter-Efficient Prompt Tuningcited by 3,795
  3. 2023QLoRA: Efficient Finetuning of Quantized LLMscited by 3,418
  4. 2019PIQA: Reasoning about Physical Commonsense in Natural Languagecited by 2,883
  5. 2018Neural Tangent Kernel: Convergence and Generalization in Neural Networkscited by 2,621
  6. 2019BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questionscited by 2,183
  7. 2020DeBERTa: Decoding-enhanced BERT with Disentangled Attentioncited by 2,181
  8. 2018Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answeringcited by 2,164
  9. 2022BLOOM: A 176B-Parameter Open-Access Multilingual Language Modelcited by 1,984
  10. 2018Know What You Don't Know: Unanswerable Questions for SQuADcited by 1,554
  11. 2018Don't Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarizationcited by 1,317
  12. 2021BitFit: Simple Parameter-efficient Fine-tuning for Transformer-based Masked Language-modelscited by 1,245
  13. 2015Teaching Machines to Read and Comprehendcited by 1,049
  14. 2018Adafactor: Adaptive Learning Rates with Sublinear Memory Costcited by 905
  15. 2022Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learningcited by 896
  16. 2021Are NLP Models really able to Solve Simple Math Word Problems?cited by 878
  17. 2021Towards a Unified View of Parameter-Efficient Transfer Learningcited by 872
  18. 2021P-Tuning v2: Prompt Tuning Can Be Comparable to Fine-tuning Universally Across Scales and Taskscited by 689
  19. 2023The Falcon Series of Open Language Modelscited by 535
  20. 2024GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projectioncited by 463
  21. 2016Solving General Arithmetic Word Problemscited by 360
  22. 2018Gradient Descent Happens in a Tiny Subspacecited by 217
  23. 2023Controlling Text-to-Image Diffusion by Orthogonal Finetuningcited by 208
  24. 2023ReLoRA: High-Rank Training Through Low-Rank Updatescited by 204
  25. 2023Progressive Prompts: Continual Learning for Language Modelscited by 197
  26. 2023S-LoRA: Serving Thousands of Concurrent LoRA Adapterscited by 152
  27. 2024Continual Learning for Large Language Models: A Surveycited by 147
  28. 2021FedPara: Low-Rank Hadamard Product for Communication-Efficient Federated Learningcited by 141
  29. 2022On the Effectiveness of Parameter-Efficient Fine-Tuningcited by 141
  30. 2020Exploring Versatile Generative Language Model Via Parameter-Efficient Transfer Learningcited by 139
  31. 2023Memory-Efficient Fine-Tuning of Compressed Large Language Models via sub-4-bit Integer Quantizationcited by 100
  32. 2020Continual Learning in Low-rank Orthogonal Subspacescited by 99
  33. 2024Chain of LoRA: Efficient Fine-tuning of Language Models via Residual Learningcited by 82
  34. 2015Fast low-rank estimation by projected gradient descent: General statistical and algorithmic guaranteescited by 70
  35. 2023Residual Prompt Tuning: Improving Prompt Tuning with Residual Reparameterizationcited by 64
  36. 2023Memory Efficient Optimizers with 4-bit Statescited by 54
  37. 2024LoRA Training in the NTK Regime has No Spurious Local Minimacited by 48
  38. 2022Dynamic Prefix-Tuning for Generative Template-based Event Extractioncited by 35

50 most cited of 61 All 61 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 3; 2025: 7; 2026: 4.

14 papers.

Sort:

By area:

  1. 2024Evaluating the Prompt Steerability of Large Language Modelscited by 27
  2. 2024Parameter-Efficient Fine-Tuning of Large Language Models using Semantic Knowledge Tuningcited by 10
  3. 2025SelfAug: Mitigating Catastrophic Forgetting in Retrieval-Augmented Generation via Distribution Self-Alignmentcited by 9
  4. 2024RoCoFT: Efficient Finetuning of Large Language Models with Row-Column Updatescited by 8
  5. 2025A Course Correction in Steerability Evaluation: Revealing Miscalibration and Side Effects in LLMscited by 7
  6. 2026Safety Alignment as Continual Learning: Mitigating the Alignment Tax via Orthogonal Gradient Projectioncited by 6
  7. 2025User Profile with Large Language Models: Construction, Updating, and Benchmarkingcited by 5
  8. 2025Predicting Through Generation: Why Generation Is Better for Predictioncited by 2
  9. 2025Zero-Shot Vision Encoder Grafting via LLM Surrogatescited by 2
  10. 2025SliceFine: The Universal Winning-Slice Hypothesis for Pretrained Networkscited by 2
  11. 2026Monkey Jump : MoE-Style PEFT for Efficient Multi-Task Learningcited by 2
  12. 2025GenFT: A Generative Parameter-Efficient Fine-Tuning Method for Pretrained Foundation Modelscited by 0
2 more
  1. 2026LiME: Lightweight Mixture of Experts for Efficient Multimodal Multi-task Learningcited by 0
  2. 2026Convex Optimization for Alignment and Preference Learning on a Single GPUcited by 0