mlsanityData to 14 Oct 2026

Delta Activations: A Representation for Finetuned Large Language Models

arXiv 2509.04442 ↗First posted 4 Sep 2025cs.LG cs.AI cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

The success of powerful open source Large Language Models (LLMs) has enabled the community to create a vast collection of post-trained models adapted to specific tasks and domains. However, navigating and understanding these models remains challenging due to inconsistent metadata and unstructured repositories. We introduce Delta Activations, a method to represent finetuned models as vector embeddings by measuring shifts in their internal activations relative to a base model. This representation allows for effective clustering by domain and task, revealing structure in the model landscape. Delta Activations also demonstrate desirable properties: it is robust across finetuning settings and exhibits an additive property when finetuning datasets are mixed. In addition, we show that Delta Activations can embed tasks via few-shot finetuning, and further explore its use for model selection and merging. We hope Delta Activations can facilitate the practice of reusing publicly available models. Code is available at https://github.com/OscarXZQ/delta_activations.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 62 of its 76 references that are in the map. First posted on arXiv 4 Sep 2025.

Builds on

62 of 76 references.

References · 62

  1. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  2. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  3. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  4. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  5. 2024Qwen2.5 Technical Reportcited by 4,942
  6. 2013Efficient Estimation of Word Representations in Vector Spacecited by 4,375
  7. 2019HellaSwag: Can a Machine Really Finish Your Sentence?cited by 4,100
  8. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
  9. 2022Scaling Instruction-Finetuned Language Modelscited by 3,024
  10. 2015Learning both Weights and Connections for Efficient Neural Networkscited by 2,031
  11. 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,998
  12. 2022GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformerscited by 1,919
38 more
  1. 2022Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Themcited by 1,807
  2. 2020What Disease does this Patient Have? A Large-scale Open Domain Question Answering Dataset from Medical Examscited by 1,356
  3. 2019PubMedQA: A Dataset for Biomedical Research Question Answeringcited by 1,288
  4. 2019A Comprehensive Survey on Transfer Learningcited by 1,160
  5. 2023SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shotcited by 1,121
  6. 2022Editing Models with Task Arithmeticcited by 1,120
  7. 2024SimPO: Simple Preference Optimization with a Reference-Free Rewardcited by 963
  8. 2023A Simple and Effective Pruning Approach for Large Language Modelscited by 800
  9. 2019Contrastive Representation Distillationcited by 748
  10. 2023MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Modelscited by 735
  11. 2023WizardCoder: Empowering Code Large Language Models with Evol-Instructcited by 733
  12. 2024RouteLLM: Learning to Route LLMs with Preference Datacited by 682
  13. 2022MedMCQA : A Large-scale Multi-Subject Multi-Choice Dataset for Medical domain Question Answeringcited by 543
  14. 2022Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Taskcited by 431
  15. 2023Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruningcited by 415
  16. 2023UltraFeedback: Boosting Language Models with Scaled AI Feedbackcited by 406
  17. 2023LoraHub: Efficient Cross-Task Generalization via Dynamic LoRA Compositioncited by 364
  18. 2023LegalBench: A Collaboratively Built Benchmark for Measuring Legal Reasoning in Large Language Modelscited by 336
  19. 2025Perception Encoder: The best visual embeddings are not at the output of the networkcited by 326
  20. 2024Massive Activations in Large Language Modelscited by 294
  21. 2024Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunitiescited by 254
  22. 2020MiniLMv2: Multi-Head Self-Attention Relation Distillation for Compressing Pretrained Transformerscited by 245
  23. 2023In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steeringcited by 245
  24. 2024Smaug: Fixing Failure Modes of Preference Optimisation with DPO-Positivecited by 241
  25. 2023Task Arithmetic in the Tangent Space: Improved Editing of Pre-Trained Modelscited by 221
  26. 2022SGPT: GPT Sentence Embeddings for Semantic Searchcited by 204
  27. 2023Large Language Model Routing with Benchmark Datasetscited by 172
  28. 2023Outlier Weighed Layerwise Sparsity (OWL): A Missing Secret Sauce for Pruning LLMs to High Sparsitycited by 165
  29. 2023Routing to the Expert: Efficient Reward-guided Ensemble of Large Language Modelscited by 145
  30. 2024OpenCoder: The Open Cookbook for Top-Tier Code Large Language Modelscited by 130
  31. 2025CodeElo: Benchmarking Competition-level Code Generation of LLMs with Human-comparable Elo Ratingscited by 115
  32. 2024Learning Dynamics of LLM Finetuningcited by 95
  33. 2023AdapterSoup: Weight Averaging to Improve Generalization of Pretrained Language Modelscited by 84
  34. 2024Towards Modular LLMs by Building and Reusing a Library of LoRAscited by 72
  35. 2024LoraRetriever: Input-Aware LoRA Retrieval and Composition for Mixed Tasks in the Wildcited by 68
  36. 2023Dynamic Sparse No Training: Training-Free Fine-tuning for Sparse LLMscited by 61
  37. 2024Localize-and-Stitch: Efficient Model Merging via Sparse Task Arithmeticcited by 50
  38. 2024Neural Network Diffusioncited by 42

50 most cited of 62 All 62 (JSON)

Built on by

Not cited yet.