mlsanityData to 14 Oct 2026

Self-Adapting Language Models

arXiv 2506.10943 ↗First posted 12 Jun 2025cs.LGcited by 84

Not on the map. Why?

Cited by 20 in Synthetic data and self-improvement for LLMs, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Large language models (LLMs) are powerful but static; they lack mechanisms to adapt their weights in response to new tasks, knowledge, or examples. We introduce Self-Adapting LLMs (SEAL), a framework that enables LLMs to self-adapt by generating their own finetuning data and update directives. Given a new input, the model produces a self-edit-a generation that may restructure the information in different ways, specify optimization hyperparameters, or invoke tools for data augmentation and gradient-based updates. Through supervised finetuning (SFT), these self-edits result in persistent weight updates, enabling lasting adaptation. To train the model to produce effective self-edits, we use a reinforcement learning loop with the downstream performance of the updated model as the reward signal. Unlike prior approaches that rely on separate adaptation modules or auxiliary networks, SEAL directly uses the model's own generation to control its adaptation process. Experiments on knowledge incorporation and few-shot generalization show that SEAL is a promising step toward language models capable of self-directed adaptation. Our website and code is available at https://jyopari.github.io/posts/seal.

Key facts. Cited by 84 papers in the map (data through 14 Oct 2026). Builds on 53 of its 73 references that are in the map. First posted on arXiv 12 Jun 2025.

Builds on

53 of 73 references.

References · 53

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2022Training language models to follow instructions with human feedbackcited by 15,971
  3. 2017Proximal Policy Optimization Algorithmscited by 15,215
  4. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  5. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  6. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  7. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,897
  8. 2016SQuAD: 100,000+ Questions for Machine Comprehension of Textcited by 4,148
  9. 2021Finetuned Language Models Are Zero-Shot Learnerscited by 3,780
  10. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,724
  11. 2022Locating and Editing Factual Associations in GPTcited by 2,691
  12. 2020Don't Stop Pretraining: Adapt Language Models to Domains and Taskscited by 1,565
38 more
  1. 2022Mass-Editing Memory in a Transformercited by 1,001
  2. 2022STaR: Bootstrapping Reasoning With Reasoningcited by 940
  3. 2013An Empirical Investigation of Catastrophic Forgetting in Gradient-Based Neural Networkscited by 722
  4. 2023Instruction Tuning with GPT-4cited by 686
  5. 2022Large Language Models Can Self-Improvecited by 664
  6. 2023RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedbackcited by 625
  7. 2024OLMo: Accelerating the Science of Language Modelscited by 613
  8. 2016RL2: Fast Reinforcement Learning via Slow Reinforcement Learningcited by 503
  9. 2021Fast Model Editing at Scalecited by 498
  10. 2023TinyStories: How Small Can Language Models Be and Still Speak Coherent English?cited by 449
  11. 2023Scaling Relationship on Learning Mathematical Reasoning with Large Language Modelscited by 427
  12. 2016Learning to reinforcement learncited by 377
  13. 2023Beyond Human Data: Scaling Self-Training for Problem-Solving with Language Modelscited by 349
  14. 2023Physics of Language Models: Part 3.1, Knowledge Storage and Extractioncited by 270
  15. 2020Modifying Memories in Transformer Modelscited by 224
  16. 2022Test-Time Training with Masked Autoencoderscited by 212
  17. 2022Will we run out of data? Limits of LLM scaling based on human-generated datacited by 195
  18. 2024Rephrasing the Web: A Recipe for Compute and Data-Efficient Language Modelingcited by 161
  19. 2023Text2Reward: Reward Shaping with Language Models for Reinforcement Learningcited by 160
  20. 2024Nemotron-CC: Transforming Common Crawl into a Refined Long-Horizon Pretraining Datasetcited by 160
  21. 2023Does Synthetic Data Generation of LLMs Help Clinical Text Mining?cited by 154
  22. 2018Meta-Reinforcement Learning of Structured Exploration Strategiescited by 130
  23. 2020Learning to Utilize Shaping Rewards: A New Approach of Reward Shapingcited by 114
  24. 2024Self-Improvement in Language Models: The Sharpening Mechanismcited by 104
  25. 2024The Surprising Effectiveness of Test-Time Training for Few-Shot Learningcited by 96
  26. 2017Meta Learning Shared Hierarchiescited by 93
  27. 2022When Should We Prefer Offline Reinforcement Learning Over Behavioral Cloning?cited by 89
  28. 2025Reward Shaping to Mitigate Reward Hacking in RLHFcited by 88
  29. 2023Language Model Self-improvement by Reinforcement Learning Contemplationcited by 75
  30. 2019Superposition of many models into onecited by 64
  31. 2023Massive Editing for Large Language Models via Meta Learningcited by 63
  32. 2024Synthetic continued pretrainingcited by 60
  33. 2025Reasoning to Learn from Latent Thoughtscited by 45
  34. 2023Meta-Learning Online Adaptation of Language Modelscited by 43
  35. 2022A Modern Self-Referential Weight Matrix That Learns to Modify Itselfcited by 41
  36. 2025On the generalization of language models from in-context learning and finetuning: a controlled studycited by 40
  37. 2024Better Synthetic Data by Retrieving and Transforming Existing Datasetscited by 36
  38. 2024CREAM: Consistency Regularized Self-Rewarding Language Modelscited by 35

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 18; 2026: 66.

84 papers.

Sort:

By area:

  1. 2025A Survey of Reinforcement Learning for Large Reasoning Modelscited by 180
  2. 2025RL's Razor: Why Online Reinforcement Learning Forgets Lesscited by 176
  3. 2025A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligencecited by 170
  4. 2025A Survey of Context Engineering for Large Language Modelscited by 126
  5. 2025Self-Questioning Language Modelscited by 50
  6. 2026A Survey of Agentic Reasoning for Large Language Models: Towards Recursively Self-Improving and Collective Agentscited by 50
  7. 2025Language Self-Play For Data-Free Trainingcited by 43
  8. 2025Self-Improving LLM Agents at Test-Timecited by 30
  9. 2026SHINE: A Scalable In-Context Hypernetwork for Mapping Context to LoRA in a Single Passcited by 25
  10. 2026Rethinking On-Policy Distillation of Large Language Models II: One Training Examplecited by 24
  11. 2026Structured Scaling of AI Discovery Across Diverse Scientific Domainscited by 21
  12. 2025LoopTool: Closing the Data-Training Loop for Robust LLM Tool Callscited by 17
38 more
  1. 2026Teaching Models to Teach Themselves: Reasoning at the Edge of Learnabilitycited by 13
  2. 2026Understanding LoRA as Knowledge Memory: An Empirical Analysiscited by 12
  3. 2026SEAL: Synergistic Co-Evolution of Agents and Learning Environmentscited by 12
  4. 2025Learning on the Job: Test-Time Curricula for Targeted Reinforcement Learningcited by 9
  5. 2026ProbeLLM: Automating Principled Diagnosis of LLM Failurescited by 9
  6. 2026Recursive Experiential-Working Memory Evolution for Long-Horizon Agent Harnessescited by 9
  7. 2026ModularRSI: Modular and Generalizable Recursive Harness Self-Improvementcited by 9
  8. 2026iReasoner: Trajectory-Aware Intrinsic Reasoning Supervision for Self-Evolving Large Multimodal Modelscited by 7
  9. 2026ScienceBuddy: Recursive-in-Recursive Self-Improvement for Interactive Scientific Agentscited by 7
  10. 2025ALAS: Autonomous Learning Agent for Self-Updating Language Modelscited by 6
  11. 2025Synthetic bootstrapped pretrainingcited by 6
  12. 2026Beyond Individual Intelligence: Surveying Collaboration, Failure Attribution, and Self-Evolution in LLM-based Multi-Agent Systemscited by 6
  13. 2025Specialization after Generalization: Towards Understanding Test-Time Training in Foundation Modelscited by 5
  14. 2026AutoSci: A Memory-Centric Agentic System for the Full Scientific Research Lifecyclecited by 5
  15. 2026Task-CoEvolve: Efficient Harness Optimization via Adaptive Validation Task Selectioncited by 5
  16. 2026Aspire: Can Models Self-Evolve from Vague Goals?cited by 5
  17. 2025No Labels, No Problem: Training Visual Reasoners with Multimodal Verifierscited by 4
  18. 2026Knowledge is Not Enough: Injecting RL Skills for Continual Adaptationcited by 4
  19. 2026Steering and Rectifying Latent Representation Manifolds in Frozen Multi-modal LLMs for Video Anomaly Detectioncited by 4
  20. 2025Maximizing Prefix-Confidence at Test-Time Efficiently Improves Mathematical Reasoningcited by 3
  21. 2026Meta-TTL: Meta-Learning Self-Improvement Policies for Language Agentscited by 3
  22. 2026Federation over Text: Insight Sharing for Multi-Agent Reasoningcited by 3
  23. 2026AIP: A Graph Representation for Learning and Governing Agent Skillscited by 3
  24. 2025Hail to the Thief: Exploring Attacks and Defenses in Decentralised GRPOcited by 2
  25. 2026A Unified Study of LoRA Variants: Taxonomy, Review, Codebase, and Empirical Evaluationcited by 2
  26. 2026Convergent World Representations and Divergent Taskscited by 2
  27. 2026Bifrost: Steering Strategic Trajectories to Bridge Contextual Gaps for Self-Improving Agentscited by 2
  28. 2026Test-Time Meta-Adaptation with Self-Synthesiscited by 2
  29. 2026Synthetic Data for any Differentiable Targetcited by 2
  30. 2026Self-Consolidating Language Models: Continual Knowledge Incorporation from Contextcited by 2
  31. 2026Language Models Need Sleep: Learning to Self-Modify and Consolidate Memoriescited by 2
  32. 2026The Path to Self-Evolving Clinical Systems: Scaling Medical Agents from Assistance to Autonomycited by 2
  33. 2026Phantom Gains: Auditing Self-Improvement Against a Measured Nullcited by 2
  34. 2026Belief-Calibrated Optimization: An Explicit World Model for Agentic Optimizationcited by 2
  35. 2026MetaRSI / RSI2: A Meta-Recursive Self-Improving System for Recursive Self-Improving Systems Themselvescited by 2
  36. 2025Self-evolving expertise in complex non-verifiable subject domains: dialogue as implicit meta-RLcited by 1
  37. 2026Iterative Closed-Loop Motion Synthesis for Scaling the Capabilities of Humanoid Controlcited by 1
  38. 2026Context Cartography: Toward Structured Governance of Contextual Space in Large Language Model Systemscited by 1

50 most cited of 84 All 84 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2024A Survey on Self-Evolution of Large Language Models
  2. 2024Training Language Models to Self-Correct via Reinforcement Learning
  3. 2026When Updating Stops Being Learning: Rethinking LLM Self-Evolution via learnable information gain
  4. 2025The Path of Self-Evolving Large Language Models: Achieving Data-Efficient Learning via Intrinsic Feedback
  5. 2025Evolving LLMs' Self-Refinement Capability via Synergistic Training-Inference Optimization
  6. 2026Learning to Self-Evolve
  7. 2026Self-Improving Large Language Models via Progressive Experience Evolution
  8. 2026On the Generalization Gap in Self-Evolving Language Model Reasoning