Pith. sign in

REVIEW 1 major objections 5 minor 195 references

Continual Learning in Transition

T0 review · 1 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read Continual learning has outgrown the parameter-centric view: this paper recasts it as capability evolution across when, where, and how.

desk verdict Useful tri-axial framework for continual learning in the LLM/agent era; the gap-based future-directions argument needs a systematic corpus check before it carries weight. read the letter →

arxiv 2608.06216 v1 pith:OBYZRM7G submitted 2026-08-06 cs.LG cs.AI

classification cs.LGcs.AI
keywords continuallearninglargelanguagemodelsagenticAIcapabilityevolutiontest-timetrainingmodelmergingagentharnesscatastrophicforgetting
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Classical continual learning treats a model that must keep learning from a stream of tasks as a problem of updating weights without forgetting what came before. This paper argues that in the era of large language models and agentic AI that picture is too narrow: capability is now acquired before, during, and after deployment; it is stored not only in parameters but also in external memory, skill libraries, and interaction protocols; and updates are driven by on-policy reinforcement and gradient-free mechanisms as much as by ordinary backpropagation. The central claim is that continual learning should be recast as continual capability evolution, with every method characterized by a When–Where–How profile instead of a position in the classical taxonomy. If true, this reframing gives researchers a common coordinate system for comparing methods across subfields and for spotting the empty regions where new work is needed.

What carries the argument

The machinery is the tri-axial framework built from three questions: When capability evolves (pre-training, post-training, or inference time), Where it is carried (parameters or the harness layer of memory, skills, and protocols), and How updates are driven (off-policy gradients, on-policy learning, or beyond-gradient mechanisms such as model merging, zeroth-order optimization, heuristic edits, and prompt evolution). The framework's work is to convert three implicit assumptions of classical continual learning into explicit variables, to locate methods in a common coordinate space, and to make dense, sparse, and empty regions visible as opportunities.

What would settle it

Conduct a systematic census of continual-learning work published between 2023 and 2026 using a keyword protocol applied to a broad index, assign every method a When–Where–How profile without reference to Figure 2, and test whether multi-axis departures from the classical corner (post-training, parameters, off-policy gradient) are growing over time and across venues. If the non-classical regions are not overrepresented or not growing, the claimed transition and the empty-region opportunities would fail to replicate.

Watch

Extended reading notes

Core claim

The paper's central claim is that the research landscape of continual learning is undergoing concurrent change along three dimensions—learning mechanism, learning timing, and locus of capability—so that classical continual learning, with its replay, regularization, architecture, and gradient families, is best read as a particular point in a larger space. In the authors' terms, classical CL corresponds to a canonical setting: post-training updates to model parameters through off-policy gradient learning. The survey assembles representative LLM-era and agent methods and assigns each a When–Where–How profile, showing that the field is moving from an isolated algorithmic problem of suppressing catastrophic forgetting in one model toward a system-level question of how capability should be organized across the model and its harness.

Load-bearing premise

The load-bearing premise is that the hand-selected methods shown in Figures 2 and 3 are representative enough of the wider literature—the paper explicitly says they are not the result of an exhaustive enumeration—so that the apparent field-level transition toward inference-time, harness-based, on-policy, and gradient-free continual learning is real and not an artifact of the authors' choices.

Editorial extensions

If this is right

  • The classical taxonomy (replay, regularization, architecture, gradient) becomes a special case rather than the whole field, and new method families can be described by their When–Where–How profiles.
  • Forgetting can no longer be measured only as parameter overwriting; it has to be tracked as retrieval decay, context overflow, and protocol mismatch, each with different timescales and reversibility.
  • Evaluation should shift from static benchmarks to trajectory-level measures of long-horizon agents, because a model that scores well on a fixed test set can still lose state and drift off goal over many interaction steps.
  • The sparse regions of the space—especially coordinated parameter-and-harness updates, active forgetting or reclamation, and on-policy inference-time learning—mark concrete research directions.
  • Continual learning for the frontier model, as opposed to domain-specific continual learning, is the priority on the path toward generally capable systems.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • One testable implication the paper leaves implicit is that the same capability can be scheduled across carriers like a memory hierarchy: frequently used, broadly shared items should be internalized into parameters, while context-specific or unstable items should stay in memory or as skills.
  • If the framework is adopted, a natural standardization would be to report every continual-learning method and benchmark with a (when, where, how) profile, making meta-analyses of the field's evolution straightforward to compute.
  • The paper's boundary claim that harness engineering is functionally but not mechanistically continual learning could be tested directly by comparing hand-maintained rule files against agent-self-evolved protocols on a shared long-horizon task set.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

1 major / 5 minor

Summary. The paper proposes a tri-axial framework for continual learning (CL) in the era of large language models and agentic AI, organized around three questions: When capability evolution occurs (pre-training, post-training, inference time), Where acquired capability is carried (parameters versus harness components such as memory, skills, and protocols), and How updates are driven (off-policy gradients, on-policy gradients, and beyond-gradient mechanisms). Classical CL is characterized as a single point in this space—post-training, parameter-level, off-policy gradient learning—and the authors survey a broad set of representative LLM-era and agentic methods through this lens. The survey argues that CL is transitioning from parameter-centric sequential training to system-level continual capability evolution, and it uses the framework to identify dense, sparse, and empty regions that motivate future research directions, including coordinated parameter–harness evolution and a priority on frontier-model CL.

Significance. If accepted, the framework provides a genuinely useful organizational device: it brings together research threads that are rarely surveyed jointly (continual pre-training, TTT, RLHF/RLVR, model merging, memory systems, skill libraries, protocol evolution) and makes their methodological relationships explicit. The paper is notably even-handed: Section 4.2 treats opposing positions on whether harness engineering already constitutes continual learning with genuine engagement, and the authors are unusually explicit about the limitations of their own method selection (Section 3.5) and of the taxonomy's boundary conditions (Section 5). The proposal to view classical CL as one point in a broader space is well supported by the many concrete examples cited. However, the forward-looking component—the identification of 'dense and empty regions' as evidence for research priorities—rests on a non-exhaustive, author-curated sample, which is a load-bearing weakness for Contribution 3 and for the concluding argument.

major comments (1)
  1. [Section 3.5 / Section 5 / Contribution 3] The paper disclaims any statistical reading of Figures 2 and 3 in Section 3.5 ('should not be interpreted as a statistical estimate of research density'), yet Contribution 3 and Section 5 rely on 'the contrast between dense and empty regions' to 'identify unfilled coordinates that offer concrete directions for future work.' This is an internal inconsistency: an unoccupied cell in a curated, non-exhaustive figure is evidence of omission, not of absence. Because the forward-looking priorities in Section 4.5 (e.g., frontier-model CL and coordinated parameter–harness evolution) are motivated in part by these perceived gaps, the paper must either provide a systematic corpus with explicit inclusion/exclusion criteria and a transparent enumeration protocol before making density-based claims, or revise Contribution 3 and Section 5 to present the empty regions as tentative hypotheses or suggestive opportunities rather than as empirical findings about the field. As written, the density-based future-directions narrative is not supported by the manuscript's own data.
minor comments (5)
  1. [Section 3.5] The sentence 'Figures 2 contains selected representative methods' has a subject–verb agreement error; it should read 'Figure 2 contains' or 'Figures 2 and 3 contain.'
  2. [Figure 2] The label 'On-pollicy' is a typo for 'On-policy,' and the entry 'Mem-�' contains a garbled glyph that should be replaced with the correct method name.
  3. [References [26] and Section 1] Reference [26] is described in the reference itself with the method label 'SDFT' in Figure 2, but the text in Section 1 calls it 'on-policy self-distillation (OPSD).' Please align the label and the acronym to avoid confusion.
  4. [Section 5 (vs. Section 3.1)] The conclusion describes the three axes as 'independent dimensions,' while Section 3.1 explicitly defines them as 'complementary perspectives' that are not mutually exclusive and allows time-varying profiles (e.g., SKILL0). Please replace 'independent' with a term such as 'complementary' or 'analytical' to avoid an inconsistency with the earlier characterization.
  5. [Section 4.2] The phrase 'We turn to that next Section (Section 4.3)' is redundant; consider 'We turn to that next (Section 4.3)' or simply 'We turn to Section 4.3.'

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the tri-axial framework is a proposed taxonomy, not a derived prediction, and the transition claim is supported by independent examples; self-citations serve only as illustrations.

full rationale

The paper's central move is to reinterpret continual learning through three questions (When, Where, How) and to place classical CL as one point in that space. This is an organizing lens, not a derived result: the axes are introduced as complementary analytical perspectives, and the paper explicitly states that a method may have multiple labels and that profiles can change over time. There is no fitted parameter, no equation whose output is also its input, and no benchmark prediction that is forced by construction. The representative-method map is admittedly non-exhaustive: 'Figures 2 contains selected representative methods and do not result from an exhaustive literature-enumeration protocol. Consequently, the number of methods displayed in a region should not be interpreted as a statistical estimate of research density, nor should an unoccupied region be treated as evidence that no relevant work exists.' That caveat protects the existence claim that classical CL is one point in a broader inhabited space. The conclusion nevertheless says the framework 'through the contrast between dense and empty regions, identifies unfilled coordinates that offer concrete directions for future work.' This is an internal tension and a genuine methodological limitation, but it is not a circular reduction: the empty regions are artifacts of the authors' curation, not fitted parameters, self-citation-derived theorems, or definitionally forced predictions. The survey also cites several works with overlapping authorship (e.g., ZeroFlow, MLLM-CTBench, CPRec, SEEKR, AlphaEdit), but these are used as examples in a representative map and do not carry the central claim; removing them would not change the taxonomy or the existence of the broader space. The load-bearing premise that a transition is underway is supported by diverse external examples and by the paper's own description of the field, not by a self-citation chain. Thus the appropriate circularity verdict is a mild score of 2 reflecting the minor self-referential flavor and the unsupported density-based forward-looking claims, without any designated circular step that reduces the derivation to its inputs.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

For a survey, the central claim rests mainly on domain assumptions about how to slice the field, not on fitted parameters or invented entities. No free parameters are used. The most important assumption is that the non-exhaustive method selection still supports the field-level transition and the sparse-region analysis.

assumptions (3)
  • domain assumption The three questions When, Where, and How are the fundamental and sufficiently independent axes for characterizing continual learning.
    The framework's usefulness depends on these being the right organizing questions; the paper provides no derivation of completeness or independence and admits boundary crossings in Section 5.
  • ad hoc to paper The representative methods shown in Figures 2 and 3 are representative of the broader literature.
    The paper states the figures are not exhaustive and counts should not be read statistically; the claimed transition and sparse-region analysis rely on this selection.
  • domain assumption Classical continual learning indeed operated under three implicit assumptions: training-stage updates, parameter carriers, and gradient-based optimization.
    This historical characterization is presented in Section 2.1 and is used as the baseline point in the taxonomy; it is an interpretation rather than a formal result.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Continual Learning in Transition." pith.science (2026). https://pith.science/paper/OBYZRM7G

@misc{pith2026260806216,
  author       = {Pith},
  title        = {Pith review of: Continual Learning in Transition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OBYZRM7G}},
  note         = {Machine review of arXiv:2608.06216}
}
read the original abstract

Classical continual learning (CL) has primarily focused on enabling models to update and retain knowledge through parameter-centric mechanisms, e.g., training strategies, architectural designs, and weight adaptation. However, emerging paradigms are reshaping the scope of CL beyond this traditional model adaptation view. For instance, on-policy learning broadens the space of update mechanisms; test-time training extends CL from the training phase to inference; and external harness components such as memory, skill libraries, and interaction protocols extend the evolutionary boundaries of model capabilities far beyond the static parameter space. Collectively, these developments indicate a transition from parameter-centric learning toward system-level adaptation. To characterize this transition, we examine the evolution of continual learning through three dimensions: When, How, and Where learning occurs. The How dimension encompasses off-policy, on-policy, and beyond-gradient optimization mechanics. The When dimension captures evolution across pre-training, post-training, and inference-time stages. The Where dimension delineates updates occurring within internal parameters versus external structural constraints. Anchored by this tri-axial framework, we systematically survey representative methods, trace the ongoing transition of continual learning, and discuss the key challenges, broader implications, and future directions arising from this paradigm shift.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

195 extracted references · 26 canonical work pages

  1. [1]

    GPT-4 technical report, 2023

    OpenAI. GPT-4 technical report, 2023. URLhttps://arxiv.org/abs/2303.08774

  2. [2]

    ChatGLM: A family of large language models from GLM-130B to GLM-4 all tools, 2024

    GLM Team. ChatGLM: A family of large language models from GLM-130B to GLM-4 all tools, 2024. URL https://arxiv.org/abs/2406.12793

  3. [3]

    Qwen technical report, 2023

    Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xiaodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, et al. Qwen technical report, 2023. URLhttps://arxiv.org/abs/2309.16609

  4. [4]

    DeepSeek-R1 incentivizes reasoning in LLMs through reinforce- ment learning.Nature, 645:633–638, 2025

    Daya Guo, Dejian Yang, Haowei Zhang, et al. DeepSeek-R1 incentivizes reasoning in LLMs through reinforce- ment learning.Nature, 645:633–638, 2025. doi: 10.1038/s41586-025-09422-z. URLhttps://doi.org/10.1038/ s41586-025-09422-z

  5. [5]

    Kimi K2.5: Visual agentic intelligence, 2026

    Kimi Team, Tongtong Bai, Yifan Bai, Yiping Bao, SH Cai, Yuan Cao, Y Charles, HS Che, Cheng Chen, Guanduo Chen, et al. Kimi K2.5: Visual agentic intelligence, 2026. URLhttps://arxiv.org/abs/2602.02276

  6. [6]

    GLM-4.5: Agentic, reasoning, and coding (ARC) foundation models, 2025

    Aohan Zeng, Xin Lv, Qinkai Zheng, Zhenyu Hou, Bin Chen, Chengxing Xie, Cunxiang Wang, Da Yin, Hao Zeng, Jiajie Zhang, et al. GLM-4.5: Agentic, reasoning, and coding (ARC) foundation models, 2025. URL https://arxiv.org/abs/2508.06471

  7. [7]

    GLM-5: From vibe coding to agentic engineering, 2026

    Aohan Zeng, Xin Lv, Zhenyu Hou, Zhengxiao Du, Qinkai Zheng, Bin Chen, Da Yin, Chendi Ge, Chengxing Xie, Cunxiang Wang, et al. GLM-5: From vibe coding to agentic engineering, 2026. URLhttps://arxiv.org/ abs/2602.15763

  8. [8]

    ReST-MCTS*: LLM self- training via process-reward-guided tree search

    Dan Zhang, Sining Zhoubian, Ziniu Hu, Yisong Yue, Yuxiao Dong, and Jie Tang. ReST-MCTS*: LLM self- training via process-reward-guided tree search. InNeurIPS, 2024

Show all 195 references
  1. [9]

    TDRM: Smooth reward models with temporal difference for LLM RL and inference, 2025

    Dan Zhang, Min Cai, Jonathan Light, Ziniu Hu, Yisong Yue, and Jie Tang. TDRM: Smooth reward models with temporal difference for LLM RL and inference, 2025. URLhttps://arxiv.org/abs/2509.15110

  2. [10]

    ReST-RL: Achieving accurate code reasoning of LLMs with optimized self-training and decoding, 2025

    Sining Zhoubian, Dan Zhang, and Jie Tang. ReST-RL: Achieving accurate code reasoning of LLMs with optimized self-training and decoding, 2025. URLhttps://arxiv.org/abs/2508.19576

  3. [11]

    Scenegenagent: Precise industrial scene generation with coding agent.arXiv preprint arXiv:2410.21909, 2024

    Xiao Xia, Dan Zhang, Zibo Liao, Zhenyu Hou, Tianrui Sun, Jing Li, Ling Fu, and Yuxiao Dong. Scenegenagent: Precise industrial scene generation with coding agent.arXiv preprint arXiv:2410.21909, 2024

  4. [12]

    MemoryLLM: Towards self-updatable large language models

    Yu Wang, Yifan Gao, Xiusi Chen, Haoming Jiang, Shiyang Li, Jingfeng Yang, Qingyu Yin, Zheng Li, Xian Li, Bing Yin, Jingbo Shang, and Julian McAuley. MemoryLLM: Towards self-updatable large language models. arXiv preprint arXiv:2402.04624, 2024

  5. [13]

    Voyager: An open-ended embodied agent with large language models, 2023

    Guanzhi Wang, Yuqi Xie, Yunfan Jiang, Ajay Mandlekar, Chaowei Xiao, Yuke Zhu, Linxi Fan, and Anima Anandkumar. Voyager: An open-ended embodied agent with large language models, 2023. URLhttps:// arxiv.org/abs/2305.16291

  6. [14]

    Reflexion: Language agents with verbal reinforcement learning, 2023

    Noah Shinn, Federico Cassano, Edward Berman, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao. Reflexion: Language agents with verbal reinforcement learning, 2023. URLhttps://arxiv.org/abs/2303. 11366

  7. [15]

    Patil, Ion Stoica, and Joseph E

    Charles Packer, Sarah Wooders, Kevin Lin, Vivian Fang, Shishir G. Patil, Ion Stoica, and Joseph E. Gonzalez. MemGPT: Towards LLMs as operating systems.arXiv preprint arXiv:2310.08560, 2023

  8. [16]

    MemoryBank: Enhancing large language models with long-term memory

    Wanjun Zhong, Lianghong Guo, Qiqi Gao, He Ye, and Yanlin Wang. MemoryBank: Enhancing large language models with long-term memory. InProceedings of the AAAI Conference on Artificial Intelligence, 2024. URL https://arxiv.org/abs/2305.10250

  9. [17]

    AgentEvolver: Towards efficient self-evolving agent system, 2025

    Yunpeng Zhai, Shuchang Tao, Cheng Chen, Anni Zou, Ziqian Chen, Qingxu Fu, Shinji Mai, Li Yu, Jiaji Deng, Zouying Cao, Zhaoyang Liu, Bolin Ding, and Jingren Zhou. AgentEvolver: Towards efficient self-evolving agent system, 2025. URLhttps://arxiv.org/abs/2511.10395

  10. [18]

    Catastrophic interference in connectionist networks: The sequential learning problem

    Michael McCloskey and Neal J Cohen. Catastrophic interference in connectionist networks: The sequential learning problem. InPsychology of learning and motivation, volume 24, pages 109–165. Elsevier, 1989

  11. [19]

    Towards continual reinforcement learning: A review and perspectives.Journal of Artificial Intelligence Research, 75:1401–1476, 2022

    Khimya Khetarpal, Matthew Riemer, Irina Rish, and Doina Precup. Towards continual reinforcement learning: A review and perspectives.Journal of Artificial Intelligence Research, 75:1401–1476, 2022. 17

  12. [20]

    A comprehensive survey of continual learning: Theory, method and application, 2024

    Liyuan Wang, Xingxing Zhang, Hang Su, and Jun Zhu. A comprehensive survey of continual learning: Theory, method and application, 2024. URLhttps://arxiv.org/abs/2302.00487

  13. [21]

    Continual learning for large language models: A survey.arXiv preprint arXiv:2402.01364, 2024

    Tongtong Wu, Linhao Luo, Yuan-Fang Li, Shirui Pan, Thuy-Trang Vu, and Gholamreza Haffari. Continual learning for large language models: A survey.arXiv preprint arXiv:2402.01364, 2024

  14. [22]

    Training language models to follow instructions with human feedback.Advances in Neural Information Processing Systems, 35:27730–27744, 2022

    Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. Training language models to follow instructions with human feedback.Advances in Neural Information Processing Systems, 35:27730...

  15. [23]

    DeepSeekMath: Pushing the limits of mathematical reasoning in open language models

    Zhihong Shao, Peiyi Wang, Qihao Zhu, Runxin Xu, Junxiao Song, Xiao Bi, Haowei Zhang, Mingchuan Zhang, YK Li, Yang Wu, et al. DeepSeekMath: Pushing the limits of mathematical reasoning in open language models. arXiv preprint arXiv:2402.03300, 2024

  16. [24]

    RL’s razor: Why online reinforcement learning forgets less,

    Idan Shenfeld, Jyothish Pari, and Pulkit Agrawal. RL’s razor: Why online reinforcement learning forgets less,

  17. [25]

    On-policy distillation of language models: Learning from self-generated mistakes

    Rishabh Agarwal, Nino Vieillard, Yongchao Zhou, Piotr Stanczyk, Sabela Ramos, Matthieu Geist, and Olivier Bachem. On-policy distillation of language models: Learning from self-generated mistakes. InInternational Conference on Learning Representations (ICLR), 2024. URLhttps://a...

  18. [26]

    Self-distillation enables continual learning,

    Idan Shenfeld, Mehul Damani, Jonas Hübotter, and Pulkit Agrawal. Self-distillation enables continual learning,

  19. [27]

    Model merging in LLMs, MLLMs, and beyond: Methods, theories, applications and opportunities.arXiv preprint arXiv:2408.07666, 2024

    Enneng Yang, Li Shen, Guibing Guo, Xingwei Wang, Xiaochun Cao, Jie Zhang, and Dacheng Tao. Model merging in LLMs, MLLMs, and beyond: Methods, theories, applications and opportunities.arXiv preprint arXiv:2408.07666, 2024

  20. [28]

    Lee, Danqi Chen, and Sanjeev Arora

    Sadhika Malladi, Tianyu Gao, Eshaan Nichani, Alex Damian, Jason D. Lee, Danqi Chen, and Sanjeev Arora. Fine-tuning language models with just forward passes. InAdvances in Neural Information Processing Systems (NeurIPS), 2023. URLhttps://arxiv.org/abs/2305.17333

  21. [29]

    Learning beyond gradients

    Jiayi Weng. Learning beyond gradients. Blog post, 2026. URLhttps://trinkle23897.github.io/ learning-beyond-gradients/. Contextual overview rather than a primary technical source

  22. [30]

    Prompt- breeder: Self-referential self-improvement via prompt evolution, 2023

    Chrisantha Fernando, Dylan Banarse, Henryk Michalewski, Simon Osindero, and Tim Rocktäschel. Prompt- breeder: Self-referential self-improvement via prompt evolution, 2023. URLhttps://arxiv.org/abs/2309. 16797

  23. [31]

    Tent: Fully test-time adaptation by entropy minimization

    Dequan Wang, Evan Shelhamer, Shaoteng Liu, Bruno Olshausen, and Trevor Darrell. Tent: Fully test-time adaptation by entropy minimization. InInternational Conference on Learning Representations, 2021. URL https://arxiv.org/abs/2006.10726

  24. [32]

    Learning to (learn at test time): RNNs with expressive hidden states

    YuSun, XinhaoLi, KaranDalal, JiaruiXu, ArjunVikram, GenghanZhang, YannDubois, XinleiChen, Xiaolong Wang, Sanmi Koyejo, Tatsunori Hashimoto, and Carlos Guestrin. Learning to (learn at test time): RNNs with expressive hidden states. InProceedings of the 42nd International Confer...

  25. [33]

    Externalization in LLM agents: A unified review of memory, skills, protocols and harness engineering, 2026

    Chenyu Zhou, Huacan Chai, Wenteng Chen, Zihan Guo, Rong Shan, Yuanyi Song, Tianyi Xu, Yingxuan Yang, Aofan Yu, Weiming Zhang, Congming Zheng, Jiachen Zhu, Zeyu Zheng, Zhuosheng Zhang, Xingyu Lou, Changwang Zhang, Zhihui Fu, Jun Wang, Weiwen Liu, Jianghao Lin, and Weinan Zhang....

  26. [34]

    An empirical investigation of catastrophic forgetting in gradient-based neural networks.arXiv preprint arXiv:1312.6211, 2013

    Ian J Goodfellow, Mehdi Mirza, Da Xiao, Aaron Courville, and Yoshua Bengio. An empirical investigation of catastrophic forgetting in gradient-based neural networks.arXiv preprint arXiv:1312.6211, 2013

  27. [35]

    icarl: Incremental classifier and representation learning

    Sylvestre-Alvise Rebuffi, Alexander Kolesnikov, Georg Sperl, and Christoph H Lampert. icarl: Incremental classifier and representation learning. InProceedings of the IEEE conference on Computer Vision and Pattern Recognition, pages 2001–2010, 2017

  28. [36]

    Experience replay for continual learning.Advances in neural information processing systems, 32, 2019

    David Rolnick, Arun Ahuja, Jonathan Schwarz, Timothy Lillicrap, and Gregory Wayne. Experience replay for continual learning.Advances in neural information processing systems, 32, 2019. 18

  29. [37]

    Infty engine: An optimization toolkit to support continual ai.GitHub repository, 2026

    Tao Feng, Wei Li, Hangjie Yuan, Liyuan Wang, Yuxiao Dong, and Minlie Huang. Infty engine: An optimization toolkit to support continual ai.GitHub repository, 2026. URLhttps://github.com/THUDM/INFTY

  30. [38]

    Gradient episodic memory for continual learning

    David Lopez-Paz and Marc’Aurelio Ranzato. Gradient episodic memory for continual learning. InAdvances in Neural Information Processing Systems, 2017

  31. [39]

    Orthogonal gradient descent for continual learning

    Mehrdad Farajtabar, Navid Azizan, Alex Mott, and Ang Li. Orthogonal gradient descent for continual learning. InInternational conference on artificial intelligence and statistics, pages 3762–3773. PMLR, 2020

  32. [40]

    Make continual learning stronger via c-flat.Advances in Neural Information Processing Systems, 37: 7608–7630, 2024

    Ang Bian, Wei Li, Hangjie Yuan, Chengrong Yu, Mang Wang, Zixiang Zhao, Aojun Lu, Pengliang Ji, and Tao Feng. Make continual learning stronger via c-flat.Advances in Neural Information Processing Systems, 37: 7608–7630, 2024

  33. [41]

    A faster path to continual learning

    Wei Li, Hangjie Yuan, Zixiang Zhao, Borui Kang, Ziwei Liu, and Tao Feng. A faster path to continual learning. CVPR, 2026

  34. [42]

    Rethinking the stability-plasticity trade-off in continual learning from an architectural perspective.ICML, 2025

    Aojun Lu, Hangjie Yuan, Tao Feng, and Yanan Sun. Rethinking the stability-plasticity trade-off in continual learning from an architectural perspective.ICML, 2025

  35. [43]

    Revisiting neural networks for continual learning: An architectural perspective.IJCAI, 2024

    Aojun Lu, Tao Feng, Hangjie Yuan, Xiaotian Song, and Yanan Sun. Revisiting neural networks for continual learning: An architectural perspective.IJCAI, 2024

  36. [44]

    Packnet: Adding multiple tasks to a single network by iterative pruning

    Arun Mallya and Svetlana Lazebnik. Packnet: Adding multiple tasks to a single network by iterative pruning. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018

  37. [45]

    Overcoming catastrophic forgetting with hard attention to the task

    Joan Serrà, Dídac Surís, Marius Miron, and Alexandros Karatzoglou. Overcoming catastrophic forgetting with hard attention to the task. InInternational Conference on Machine Learning, 2018

  38. [46]

    Rusu, Neil C

    Andrei A. Rusu, Neil C. Rabinowitz, Guillaume Desjardins, Hubert Soyer, James Kirkpatrick, Koray Kavukcuoglu, Razvan Pascanu, and Raia Hadsell. Progressive neural networks.arXiv preprint arXiv:1606.04671, 2016

  39. [47]

    Overcoming catastrophic forgetting in incremental object detection via elastic response distillation

    Tao Feng, Mang Wang, and Hangjie Yuan. Overcoming catastrophic forgetting in incremental object detection via elastic response distillation. InCVPR, pages 9417–9426. IEEE, 2022

  40. [48]

    Rusu, Kieran Milan, John Quan, Tiago Ramalho, Agnieszka Grabska-Barwinska, Demis Hassabis, Claudia Clopath, DharshanKumaran, andRaiaHadsell

    James Kirkpatrick, Razvan Pascanu, Neil Rabinowitz, Joel Veness, Guillaume Desjardins, Andrei A. Rusu, Kieran Milan, John Quan, Tiago Ramalho, Agnieszka Grabska-Barwinska, Demis Hassabis, Claudia Clopath, DharshanKumaran, andRaiaHadsell. Overcomingcatastrophicforgettinginneura...

  41. [49]

    Continual learning through synaptic intelligence

    Friedemann Zenke, Ben Poole, and Surya Ganguli. Continual learning through synaptic intelligence. InInter- national conference on machine learning, pages 3987–3995. PMLR, 2017

  42. [50]

    Memory aware synapses: Learning what (not) to forget, 2018

    Rahaf Aljundi, Francesca Babiloni, Mohamed Elhoseiny, Marcus Rohrbach, and Tinne Tuytelaars. Memory aware synapses: Learning what (not) to forget, 2018. URLhttps://arxiv.org/abs/1711.09601

  43. [51]

    Learning without forgetting.IEEE transactions on pattern analysis and machine intelligence, 40(12):2935–2947, 2017

    Zhizhong Li and Derek Hoiem. Learning without forgetting.IEEE transactions on pattern analysis and machine intelligence, 40(12):2935–2947, 2017

  44. [52]

    Richter, Quentin Anthony, Timothée Lesort, Eugene Belilovsky, and Irina Rish

    Adam Ibrahim, Benjamin Thérien, Kshitij Gupta, Mats L. Richter, Quentin Anthony, Timothée Lesort, Eugene Belilovsky, and Irina Rish. Simple and scalable strategies to continually pre-train large language models, 2024. URLhttps://arxiv.org/abs/2403.08763. Method label in Figure...

  45. [53]

    Richter, Quentin Anthony, Eugene Belilovsky, Irina Rish, and Timothée Lesort

    Kshitij Gupta, Benjamin Thérien, Adam Ibrahim, Mats L. Richter, Quentin Anthony, Eugene Belilovsky, Irina Rish, and Timothée Lesort. Continual pre-training of large language models: How to (re)warm your model?,

  46. [54]

    Towards continual knowledge learning of language models

    Joel Jang, Seonghyeon Ye, Sohee Yang, Joongbo Shin, Janghoon Han, Gyeonghun Kim, Stanley Jungkyu Choi, and Minjoon Seo. Towards continual knowledge learning of language models. InInternational Conference on Learning Representations (ICLR), 2022

  47. [55]

    ELLE: Efficient lifelong pre-training for emerging data

    Yujia Qin, Jiajie Zhang, Yankai Lin, Zhiyuan Liu, Peng Li, Maosong Sun, and Jie Zhou. ELLE: Efficient lifelong pre-training for emerging data. InFindings of the Association for Computational Linguistics: ACL 2022, 2022. 19

  48. [56]

    TimeLMs: Diachronic language models from twitter

    Daniel Loureiro, Francesco Barbieri, Leonardo Neves, Luis Espinosa Anke, and Jose Camacho-Collados. TimeLMs: Diachronic language models from twitter. InProceedings of ACL 2022: System Demonstrations, 2022

  49. [57]

    Large language model empowered recommendation meets all-domain continual pre-training.IEEE Transactions on Knowledge and Data Engineering, pages 1–14, 2026

    Haokai Ma, Yunshan Ma, Ruobing Xie, Lei Meng, Jialie Shen, Xingwu Sun, Zhanhui Kang, and Tat-Seng Chua. Large language model empowered recommendation meets all-domain continual pre-training.IEEE Transactions on Knowledge and Data Engineering, pages 1–14, 2026. doi: 10.1109/TKD...

  50. [58]

    End- to-end test-time training for long context.arXiv preprint arXiv:2512.23675, 2025

    Arnuv Tandon, Karan Dalal, Xinhao Li, Daniel Koceja, Marcel Rød, Sam Buchanan, Xiaolong Wang, Jure Leskovec, Sanmi Koyejo, Tatsunori Hashimoto, Carlos Guestrin, Jed McCaleb, Yejin Choi, and Yu Sun. End- to-end test-time training for long context.arXiv preprint arXiv:2512.23675, 2025

  51. [59]

    Titans: Learning to memorize at test time, 2025

    Ali Behrouz, Peilin Zhong, and Vahab Mirrokni. Titans: Learning to memorize at test time, 2025

  52. [60]

    Orthogonal subspace learning for language model continual learning

    Xiao Wang, Tianze Chen, Qiming Ge, Han Xia, Rong Bao, Rui Zheng, Qi Zhang, Tao Gui, and Xuan-Jing Huang. Orthogonal subspace learning for language model continual learning. InFindings of the Association for Computational Linguistics: EMNLP 2023, pages 10658–10671, 2023

  53. [61]

    Progres- sive prompts: Continual learning for language models

    Anastasia Razdaibiedina, Yuning Mao, Rui Hou, Madian Khabsa, Mike Lewis, and Amjad Almahairi. Progres- sive prompts: Continual learning for language models. InThe Eleventh International Conference on Learning Representations, 2023

  54. [62]

    LoRAMoE: Alleviating world knowledge forgetting in large language models via MoE-style plugin

    Shihan Dou, Enyu Zhou, Yan Liu, Songyang Gao, Wei Shen, et al. LoRAMoE: Alleviating world knowledge forgetting in large language models via MoE-style plugin. InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics, 2024

  55. [63]

    SLIM: Let LLMs learn more and forget less with soft LoRA and identity mixture

    Jiayi Han, Liang Du, Hongwei Du, Xiangguo Zhou, Yiwen Wu, Yuanfang Zhang, Weibo Zheng, and Donghong Han. SLIM: Let LLMs learn more and forget less with soft LoRA and identity mixture. InProceedings of the 2025 Conference of the Nations of the Americas Chapter of the Associatio...

  56. [64]

    SAPT: A shared attention framework for parameter-efficient continual learning of large language models

    Weixiang Zhao, Shilong Wang, Yulin Hu, Yanyan Zhao, Bing Qin, Xuanyu Zhang, Qing Yang, Dongliang Xu, and Wanxiang Che. SAPT: A shared attention framework for parameter-efficient continual learning of large language models. InProceedings of the 62nd Annual Meeting of the Associ...

  57. [65]

    Rehearsal-free modular and compositional continual learning for language models

    Mingyang Wang, Heike Adel, Lukas Lange, Jannik Strötgen, and Hinrich Schütze. Rehearsal-free modular and compositional continual learning for language models. InProceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics (N...

  58. [66]

    InsCL: A data-efficient continual learning paradigm for fine-tuning large language models with instructions

    Yifan Wang, Yafei Liu, Chufan Shi, Haoling Li, Chen Chen, Haonan Lu, and Yujiu Yang. InsCL: A data-efficient continual learning paradigm for fine-tuning large language models with instructions. InProceedings of the 2024 Conference of the North American Chapter of the Associati...

  59. [67]

    Mitigating catastrophic forgetting in large language models with self-synthesized rehearsal

    Jianheng Huang, Leyang Cui, Ante Wang, Chengyi Yang, Xinting Liao, Linfeng Song, Junfeng Yao, and Jinsong Su. Mitigating catastrophic forgetting in large language models with self-synthesized rehearsal. InProceedings of the 62nd Annual Meeting of the Association for Computatio...

  60. [68]

    Seekr: Selective attention- guided knowledge retention for continual learning of large language models.arXiv preprint arXiv:2411.06171, 2024

    Jinghan He, Haiyun Guo, Kuan Zhu, Zihan Zhao, Ming Tang, and Jinqiao Wang. Seekr: Selective attention- guided knowledge retention for continual learning of large language models.arXiv preprint arXiv:2411.06171, 2024

  61. [69]

    AlphaEdit: Null-space constrained knowledge editing for language models

    Junfeng Fang, Houcheng Jiang, Kun Wang, Yunshan Ma, Shi Jie, Xiang Wang, Xiangnan He, and Tat-Seng Chua. AlphaEdit: Null-space constrained knowledge editing for language models. InInternational Conference on Learning Representations (ICLR), 2025

  62. [70]

    Norm anchors make model edits last, 2026

    Mingda Liu, Zhenghan Zhu, Ze’an Miao, and Katsuki Fujisawa. Norm anchors make model edits last, 2026

  63. [71]

    Yu, and Xiao-Ming Wu

    Yujie Feng, Hao Wang, Jian Li, Xu Chu, Zhaolu Kang, Yiran Liu, Yasha Wang, Philip S. Yu, and Xiao-Ming Wu. FOREVER: Forgetting curve-inspired memory replay for language model continual learning, 2026

  64. [72]

    Dynamic cross-modal prompt generation for multimodal continual instruction tuning, 2026

    Tao Hu and Da-Wei Zhou. Dynamic cross-modal prompt generation for multimodal continual instruction tuning, 2026. 20

  65. [73]

    CRAM: Centroid-routing and adaptive MoE for multimodal continual instruction tuning, 2026

    Jun-Tao Tang, Zhen-Hao Xie, Yu-Cheng Shi, and Da-Wei Zhou. CRAM: Centroid-routing and adaptive MoE for multimodal continual instruction tuning, 2026

  66. [74]

    Ziqi Wang, Chang Che, Qi Wang, Hui Ma, Zenglin Shi, Cees G. M. Snoek, and Meng Wang. Harmonious parameter adaptation in continual visual instruction tuning for safety-aligned MLLMs, 2025

  67. [75]

    Hidden forgetting in continual multimodal learning: When accuracy survives but grounding fails, 2026

    Qianyu Chen, Canran Xiao, and Runxuan Tang. Hidden forgetting in continual multimodal learning: When accuracy survives but grounding fails, 2026

  68. [76]

    Rethinking continual experience internalization for self-evolving LLM agents, 2026

    Jingwen Chen, Wenkai Yang, Shengda Fan, Wenbo Nie, Chenxing Sun, Shaodong Zheng, Yangen Hu, Lu Pan, Ke Zeng, and Yankai Lin. Rethinking continual experience internalization for self-evolving LLM agents, 2026

  69. [77]

    Language models need sleep: Learning to self-modify and consolidate memories, 2026

    Ali Behrouz, Farnoosh Hashemi, and Vahab Mirrokni. Language models need sleep: Learning to self-modify and consolidate memories, 2026

  70. [78]

    Peam: Parametric embodied agent memory through contrastive internalization of experience in minecraft, 2026

    Yuchen Guo, Junli Gong, Weicheng Wang, Hongmin Cai, Yiu ming Cheung, and Weifeng Su. Peam: Parametric embodied agent memory through contrastive internalization of experience in minecraft, 2026

  71. [79]

    Evolving-rl: End-to-end optimization of experience-driven self-evolving capability within agents, 2026

    Zhiyuan Fan, Wenwei Jin, Feng Zhang, Bin Li, Yihong Dong, Yao Hu, and Jiawei Li. Evolving-rl: End-to-end optimization of experience-driven self-evolving capability within agents, 2026

  72. [80]

    A-MEM: Agentic memory for LLM agents, 2025

    Wujiang Xu, Zujie Liang, Kai Mei, Hang Gao, Juntao Tan, and Yongfeng Zhang. A-MEM: Agentic memory for LLM agents, 2025. URLhttps://arxiv.org/abs/2502.12110

  73. [81]

    HippoRAG: Neurobiologically inspired long-term memory for large language models

    Bernal Jiménez Gutiérrez, Yiheng Shu, Yu Gu, Michihiro Yasunaga, and Yu Su. HippoRAG: Neurobiologically inspired long-term memory for large language models. InThe Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

  74. [82]

    ExpeL: LLM agents are experiential learners

    Andrew Zhao, Daniel Huang, Quentin Xu, Matthieu Lin, Yong-Jin Liu, and Gao Huang. ExpeL: LLM agents are experiential learners. InProceedings of the AAAI Conference on Artificial Intelligence, 2024

  75. [83]

    Agent workflow memory

    Zora Zhiruo Wang, Jiayuan Mao, Daniel Fried, and Graham Neubig. Agent workflow memory. InInternational Conference on Machine Learning (ICML), 2025. Online setting shown as AWM-online in Figure 2

  76. [84]

    Mem0: Building production- ready AI agents with scalable long-term memory, 2025

    Prateek Chhikara, Dev Khant, Saket Aryan, Taranjeet Singh, and Deshraj Yadav. Mem0: Building production- ready AI agents with scalable long-term memory, 2025

  77. [85]

    Towards scalable lifelong knowledge editing with selective knowledge suppression, 2026

    Dahyun Jung, Jaewook Lee, and Heuiseok Lim. Towards scalable lifelong knowledge editing with selective knowledge suppression, 2026

  78. [86]

    Forget to improve: On-device LLM-agent continual learning via budget-curated memory, 2026

    Beining Wu, Zihao Ding, Jun Huang, and Yanxiao Zhao. Forget to improve: On-device LLM-agent continual learning via budget-curated memory, 2026

  79. [87]

    Collaborative multi-agent test-time reinforcement learning for reasoning, 2026

    Zhiyuan Hu, Yunhai Hu, Juncheng Liu, Shuyue Stella Li, Yucheng Wang, Zhen Xu, Xinxing Xu, See-Kiong Ng, Anh Tuan Luu, Bryan Hooi, Cynthia Breazeal, and Hae Won Park. Collaborative multi-agent test-time reinforcement learning for reasoning, 2026

  80. [88]

    Aging with GRACE: Lifelong model editing with discrete key-value adaptors

    Tom Hartvigsen, Swami Sankaranarayanan, Hamid Palangi, Yoon Kim, and Marzyeh Ghassemi. Aging with GRACE: Lifelong model editing with discrete key-value adaptors. InAdvances in Neural Information Processing Systems (NeurIPS), 2023

  81. [89]

    WISE: Rethinking the knowledge memory for lifelong model editing of large language models

    Peng Wang, Zexi Li, Ningyu Zhang, Ziwen Xu, Yunzhi Yao, Yong Jiang, Pengjun Xie, Fei Huang, and Huajun Chen. WISE: Rethinking the knowledge memory for lifelong model editing of large language models. InAdvances in Neural Information Processing Systems (NeurIPS), 2024

  82. [90]

    MemRL: Self-evolving agents via runtime reinforcement learning on episodic memory, 2026

    Shengtao Zhang, Jiaqian Wang, Ruiwen Zhou, Junwei Liao, Yuchen Feng, Weinan Zhang, Ying Wen, Zhiyu Li, Feiyu Xiong, Yutao Qi, Bo Tang, and Muning Wen. MemRL: Self-evolving agents via runtime reinforcement learning on episodic memory, 2026. URLhttps://arxiv.org/abs/2601.03192

  83. [91]

    Pan, Hinrich Schütze, Volker Tresp, and Yunpu Ma

    Sikuan Yan, Xiufeng Yang, Zuchao Huang, Ercong Nie, Zifeng Ding, Zonggen Li, Xiaowen Ma, Jinhe Bi, Kristian Kersting, Jeff Z. Pan, Hinrich Schütze, Volker Tresp, and Yunpu Ma. Memory-r1: Enhancing large language model agents to manage and utilize memories via reinforcement lea...

  84. [92]

    Mem- α: Learning memory construction via reinforcement learning, 2025

    Yu Wang, Ryuichi Takanobu, Zhiqi Liang, Yuzhen Mao, Yuanzhe Hu, Julian McAuley, and Xiaojian Wu. Mem- α: Learning memory construction via reinforcement learning, 2025. 21

  85. [93]

    Memory-R2: Fair credit assignment for long-horizon memory-augmented LLM agents, 2026

    Sikuan Yan, Ahmed Bahloul, Ercong Nie, Susanna Schwarzmann, Riccardo Trivisonno, Volker Tresp, and Yunpu Ma. Memory-R2: Fair credit assignment for long-horizon memory-augmented LLM agents, 2026

  86. [94]

    MemBuilder: Reinforcing LLMs for long-term memory construction via attributed dense rewards, 2026

    Zhiyu Shen, Ziming Wu, Fuming Lai, Shaobing Lian, and Yanghui Rao. MemBuilder: Reinforcing LLMs for long-term memory construction via attributed dense rewards, 2026

  87. [95]

    Memq: Integrating q-learning into self-evolving memory agents over provenance dags, 2026

    Junwei Liao, Haoting Shi, Ruiwen Zhou, Jiaqian Wang, Shengtao Zhang, Wei Zhang, Ying Wen, Zhiyu Li, Feiyu Xiong, Bo Tang, Weinan Zhang, and Muning Wen. Memq: Integrating q-learning into self-evolving memory agents over provenance dags, 2026

  88. [96]

    Marginal advantage accumulation for memory-driven agent self-evolution, 2026

    Mingyu Yang, Keye Zheng, Congchao Cheng, Yujie Liu, Xingkang Lu, Fan Jiang, and Yefei Zheng. Marginal advantage accumulation for memory-driven agent self-evolution, 2026

  89. [97]

    Just-in-time reinforce- ment learning: Continual learning in LLM agents without gradient updates, 2026

    Yibo Li, Zijie Lin, Ailin Deng, Xuan Zhang, Yufei He, Shuo Ji, Tri Cao, and Bryan Hooi. Just-in-time reinforce- ment learning: Continual learning in LLM agents without gradient updates, 2026

  90. [98]

    Le, Samira Daruki, Xiangru Tang, Vishy Tirumalashetty, George Lee, Mahsan Rofouei, Hangfei Lin, Jiawei Han, Chen-Yu Lee, and Tomas Pfister

    Siru Ouyang, Jun Yan, I-Hung Hsu, Yanfei Chen, Ke Jiang, Zifeng Wang, Rujun Han, Long T. Le, Samira Daruki, Xiangru Tang, Vishy Tirumalashetty, George Lee, Mahsan Rofouei, Hangfei Lin, Jiawei Han, Chen-Yu Lee, and Tomas Pfister. Reasoningbank: Scaling agent self-evolving with ...

  91. [99]

    Learning on the job: An experience-driven self-evolving agent for long-horizon tasks, 2025

    Cheng Yang, Xuemeng Yang, Licheng Wen, Daocheng Fu, Jianbiao Mei, Rong Wu, Pinlong Cai, Yufan Shen, Nianchen Deng, Botian Shi, Yu Qiao, and Haifeng Li. Learning on the job: An experience-driven self-evolving agent for long-horizon tasks, 2025

  92. [100]

    Exg: Self-evolving agents with experience graphs, 2026

    Yuxin Jin, Siyuan Zhang, Hanchen Wang, Lu Qin, Ying Zhang, and Wenjie Zhang. Exg: Self-evolving agents with experience graphs, 2026

  93. [101]

    The past is prologue: A plug-in controller for selective updates in sequentially evolving LLM memory, 2026

    Zihan Chen, Songwei Dong, Chengshuai Shi, Peng Wang, Song Wang, Cong Shen, and Jundong Li. The past is prologue: A plug-in controller for selective updates in sequentially evolving LLM memory, 2026

  94. [102]

    Simple- Mem: Efficient lifelong memory for LLM agents, 2026

    Jiaqi Liu, Yaofeng Su, Peng Xia, Siwei Han, Zeyu Zheng, Cihang Xie, Mingyu Ding, and Huaxiu Yao. Simple- Mem: Efficient lifelong memory for LLM agents, 2026

  95. [103]

    Automem: Automated learning of memory as a cognitive skill, 2026

    Shengguang Wu, Hao Zhu, Yuhui Zhang, Xiaohan Wang, and Serena Yeung-Levy. Automem: Automated learning of memory as a cognitive skill, 2026

  96. [104]

    Autoagent: Evolving cognition and elastic memory orchestration for adaptive agents, 2026

    Xiaoxing Wang, Ning Liao, Shikun Wei, Chen Tang, and Feiyu Xiong. Autoagent: Evolving cognition and elastic memory orchestration for adaptive agents, 2026

  97. [105]

    Toward generalist autonomous research via hypothesis-tree refinement, 2026

    Jiajie Jin, Yuyang Hu, Kai Qiu, Qi Dai, Chong Luo, Guanting Dong, Xiaoxi Li, Tong Zhao, Xiaolong Ma, Gongrui Zhang, Zhirong Wu, Bei Liu, Zhengyuan Yang, Linjie Li, Lijuan Wang, Hongjin Qian, Yutao Zhu, and Zhicheng Dou. Toward generalist autonomous research via hypothesis-tree...

  98. [106]

    Evoclinician: A self-evolving agent for multi-turn medical diagnosis via test-time evolutionary learning, 2026

    Yufei He, Juncheng Liu, Zhiyuan Hu, Yulin Chen, Yue Liu, Yuan Sui, Yibo Li, Nuo Chen, Jun Hu, Bryan Hooi, Xinxing Xu, and Jiang Bian. Evoclinician: A self-evolving agent for multi-turn medical diagnosis via test-time evolutionary learning, 2026

  99. [107]

    Scaling self-evolving agents via parametric memory, 2026

    Tao Ren, Weiyao Luo, Hui Yang, Rongzhi Zhu, Xiang Huang, Yuchuan Wu, Bingxue Chou, Jieping Ye, Jiafeng Liang, Yongbin Li, and Yijie Peng. Scaling self-evolving agents via parametric memory, 2026. URLhttps: //arxiv.org/abs/2606.04536. Method label in Figure 2: TMEM

  100. [108]

    TroVE: Inducing verifiable and efficient toolboxes for solving programmatic tasks

    Zhiruo Wang, Daniel Fried, and Graham Neubig. TroVE: Inducing verifiable and efficient toolboxes for solving programmatic tasks. InInternational Conference on Machine Learning (ICML), 2024

  101. [109]

    SkillRL: Evolving agents via recursive skill- augmented reinforcement learning, 2026

    Peng Xia, Jianwen Chen, Hanyang Wang, Jiaqi Liu, Kaide Zeng, Yu Wang, Siwei Han, Yiyang Zhou, Xujiang Zhao, Haifeng Chen, Zeyu Zheng, Cihang Xie, and Huaxiu Yao. SkillRL: Evolving agents via recursive skill- augmented reinforcement learning, 2026. URLhttps://arxiv.org/abs/2602.08234

  102. [110]

    Reinforcement learning for self-improving agent with skill library.arXiv preprint arXiv:2512.17102, 2025

    Jiongxiao Wang, Qiaojing Yan, Yawei Wang, Yijun Tian, Soumya Smruti Mishra, Zhichao Xu, Megha Gandhi, and Panpan Xu. Reinforcement learning for self-improving agent with skill library.arXiv preprint arXiv:2512.17102, 2025

  103. [111]

    AutoSkill: Experience-driven lifelong learning via skill self-evolution, 2026

    Yutao Yang, Junsong Li, Qianjun Pan, Bihao Zhan, Yuxuan Cai, Lin Du, Jie Zhou, Kai Chen, Qin Chen, Xin Li, Bo Zhang, and Liang He. AutoSkill: Experience-driven lifelong learning via skill self-evolution, 2026

  104. [112]

    Reskill: Reconciling skill creation with policy optimization in agentic rl, 2026

    Zelin He, Haotian Lin, Boran Han, Wei Zhu, Haoyang Fang, Bernie Wang, Xuan Zhu, Runze Li, and Matthew Reimherr. Reskill: Reconciling skill creation with policy optimization in agentic rl, 2026. 22

  105. [113]

    Arise: Agent reasoning with intrinsic skill evolution in hierarchical reinforcement learning, 2026

    Yu Li, Rui Miao, Zhengling Qi, and Tian Lan. Arise: Agent reasoning with intrinsic skill evolution in hierarchical reinforcement learning, 2026

  106. [114]

    Skillgraph: Skill- augmented reinforcement learning for agents via evolving skill graphs, 2026

    Xiaoyuan Li, Moxin Li, Keqin Bao, Yubo Ma, Wenjie Wang, Dayiheng Liu, and Fuli Feng. Skillgraph: Skill- augmented reinforcement learning for agents via evolving skill graphs, 2026

  107. [115]

    Skill1: Unified evolution of skill-augmented agents via reinforcement learning, 2026

    Yaorui Shi, Yuxin Chen, Zhengxi Lu, Yuchun Miao, Shugui Liu, Qi Gu, Xunliang Cai, Xiang Wang, and An Zhang. Skill1: Unified evolution of skill-augmented agents via reinforcement learning, 2026

  108. [116]

    Skillcomposer: Learning to evolve agent skills for specification and generalization, 2026

    Qi Zhang, Zhaopeng Feng, Xiaonan Shi, Xiaomeng Hu, Chu Liu, Pengjun Xie, Xiaobin Wang, Jieping Ye, Bryan Hooi, Haobo Wang, and Junbo Zhao. Skillcomposer: Learning to evolve agent skills for specification and generalization, 2026

  109. [117]

    From atomic actions to standard operating procedures: Iterative tool optimization for self-evolving LLM agents, 2026

    Haipeng Ding, Yuexiang Xie, Zhewei Wei, Yaliang Li, and Bolin Ding. From atomic actions to standard operating procedures: Iterative tool optimization for self-evolving LLM agents, 2026

  110. [118]

    Agentfactory: A self-evolving framework through executable subagent accumulation and reuse, 2026

    Zhang Zhang, Shuqi Lu, Hongjin Qian, Di He, and Zheng Liu. Agentfactory: A self-evolving framework through executable subagent accumulation and reuse, 2026

  111. [119]

    Tool-making and self-evolving LLM agents in low-latency systems, 2026

    Kalle Kujanpää, Ning Liu, Shahnawaz Alam, Yeshwanth Reddy Sura, Tianyu Yang, Kristina Klinkner, and Shervin Malmasi. Tool-making and self-evolving LLM agents in low-latency systems, 2026

  112. [120]

    Dreamprover: Evolving transferable lemma libraries via a wake-sleep theorem-proving agent, 2026

    Youyuan Zhang, Jialiang Sun, Hangrui Bi, Chuqin Geng, Wenjie Ma, Zhaoyu Li, and Xujie Si. Dreamprover: Evolving transferable lemma libraries via a wake-sleep theorem-proving agent, 2026

  113. [121]

    Ratchet: A minimal hygiene recipe for self-evolving LLM agents, 2026

    Xing Zhang, Yanwei Cui, Guanghui Wang, Ziyuan Li, Wei Qiu, Bing Zhu, and Peiyang He. Ratchet: A minimal hygiene recipe for self-evolving LLM agents, 2026

  114. [122]

    Fatemi, Xiaolong Jin, Zora Zhiruo Wang, Apurva Gandhi, Yueqi Song, Yu Gu, Jayanth Srinivasa, Gaowen Liu, Graham Neubig, and Yu Su

    Boyuan Zheng, Michael Y. Fatemi, Xiaolong Jin, Zora Zhiruo Wang, Apurva Gandhi, Yueqi Song, Yu Gu, Jayanth Srinivasa, Gaowen Liu, Graham Neubig, and Yu Su. SkillWeaver: Web agents can self-improve by discovering and honing skills, 2025. URLhttps://arxiv.org/abs/2504.07079

  115. [123]

    Inducing programmatic skills for agentic tasks, 2025

    Zora Zhiruo Wang, Apurva Gandhi, Graham Neubig, and Daniel Fried. Inducing programmatic skills for agentic tasks, 2025

  116. [124]

    Gonzalez, and Bin Cui

    Ling Yang, Zhaochen Yu, Tianjun Zhang, Shiyi Cao, Minkai Xu, Wentao Zhang, Joseph E. Gonzalez, and Bin Cui. Buffer of thoughts: Thought-augmented reasoning with large language models, 2024

  117. [125]

    Alita: Generalist agent enabling scalable agentic reasoning with minimal predefinition and maximal self-evolution, 2025

    Jiahao Qiu, Xuan Qi, Tongcheng Zhang, Xinzhe Juan, Jiacheng Guo, Yifu Lu, Yimin Wang, Zixin Yao, Qihan Ren, Xun Jiang, Xing Zhou, Dongrui Liu, Ling Yang, Yue Wu, Kaixuan Huang, Shilong Liu, Hongru Wang, and Mengdi Wang. Alita: Generalist agent enabling scalable agentic reasoni...

  118. [126]

    Memp: Exploring agent procedural memory, 2025

    Runnan Fang, Yuan Liang, Xiaobin Wang, Jialong Wu, Shuofei Qiao, Pengjun Xie, Fei Huang, Huajun Chen, and Ningyu Zhang. Memp: Exploring agent procedural memory, 2025

  119. [127]

    Skill-Pro: Learning reusable skills from experience via non-parametric PPO for LLM agents, 2026

    Qirui Mi, Zhijian Ma, Mengyue Yang, Haoxuan Li, Yisen Wang, Haifeng Zhang, and Jun Wang. Skill-Pro: Learning reusable skills from experience via non-parametric PPO for LLM agents, 2026

  120. [128]

    Muse-autoskill: Self-evolving agents via skill creation, memory, management, and evaluation, 2026

    Huawei Lin, Peng Li, Jie Song, Fuxin Jiang, and Tieying Zhang. Muse-autoskill: Self-evolving agents via skill creation, memory, management, and evaluation, 2026

  121. [129]

    AutoRefine: From trajectories to reusable expertise for continual LLM agent refinement, 2026

    Libin Qiu, Zhirong Gao, Junfu Chen, Yuhang Ye, Weizhi Huang, Xiaobo Xue, Wenkai Qiu, and Shuo Tang. AutoRefine: From trajectories to reusable expertise for continual LLM agent refinement, 2026

  122. [130]

    Mem2evolve: Towards self-evolving agents via co-evolutionary capability expansion and experience distillation, 2026

    Zihao Cheng, Zeming Liu, Yingyu Shan, Xinyi Wang, Xiangrong Zhu, Yunpu Ma, Hongru Wang, Yuhang Guo, Wei Lin, and Yunhong Wang. Mem2evolve: Towards self-evolving agents via co-evolutionary capability expansion and experience distillation, 2026

  123. [131]

    Skillforge: Forging domain-specific, self-evolving agent skills in cloud technical support, 2026

    Xingyan Liu, Xiyue Luo, Linyu Li, Ganghong Huang, Jianfeng Liu, and Honglin Qiao. Skillforge: Forging domain-specific, self-evolving agent skills in cloud technical support, 2026

  124. [132]

    Federatedskill: Federated learning for agentic skill evolution, 2026

    Jingbo Yang, Guanyu Yao, Yang Zhang, Ramana Rao Kompella, Gaowen Liu, and Shiyu Chang. Federatedskill: Federated learning for agentic skill evolution, 2026

  125. [133]

    Mem- skill: Learning and evolving memory skills for self-evolving agents, 2026

    Haozhen Zhang, Quanyu Long, Jianzhu Bao, Tao Feng, Weizhi Zhang, Haodong Yue, and Wenya Wang. Mem- skill: Learning and evolving memory skills for self-evolving agents, 2026. 23

  126. [134]

    Agent0: Unleashing self-evolving agents from zero data via tool-integrated reasoning, 2025

    Peng Xia, Kaide Zeng, Jiaqi Liu, Can Qin, Fang Wu, Yiyang Zhou, Caiming Xiong, and Huaxiu Yao. Agent0: Unleashing self-evolving agents from zero data via tool-integrated reasoning, 2025. URLhttps://arxiv.org/ abs/2511.16043

  127. [135]

    Learning while acting: A skill-enhanced test-time co-evolution framework for online lifelong learning agents, 2026

    Bo Mao, Jie Zhou, Yutao Yang, Xin Li, Xian Wei, Qin Chen, Xingjiao Wu, and Liang He. Learning while acting: A skill-enhanced test-time co-evolution framework for online lifelong learning agents, 2026

  128. [136]

    SKILL0: In-context agentic reinforcement learning for skill internalization, 2026

    Zhengxi Lu, Zhiyuan Yao, Jinyang Wu, Chengcheng Han, Qi Gu, Xunliang Cai, Weiming Lu, Jun Xiao, Yueting Zhuang, and Yongliang Shen. SKILL0: In-context agentic reinforcement learning for skill internalization, 2026. URLhttps://arxiv.org/abs/2604.02268

  129. [137]

    Guanyu Jiang, Zhaochen Su, Xiaoye Qu, and Yi R. Fung. Xskill: Continual learning from experience and skills in multimodal agents, 2026

  130. [138]

    Agrawal, Shangyin Tan, Dilara Soylu, Noah Ziems, Rishi Khare, Krista Opsahl-Ong, Arnav Singhvi, Herumb Shandilya, Michael J

    Lakshya A. Agrawal, Shangyin Tan, Dilara Soylu, Noah Ziems, Rishi Khare, Krista Opsahl-Ong, Arnav Singhvi, Herumb Shandilya, Michael J. Ryan, Meng Jiang, Christopher Potts, Koushik Sen, Alexandros G. Dimakis, Ion Stoica, Dan Klein, Matei Zaharia, and Omar Khattab. GEPA: Reflec...

  131. [139]

    Optima: Optimizing effectiveness and efficiency for LLM-based multi-agent systems, 2024

    Weize Chen, Jiarui Yuan, Chen Qian, Cheng Yang, Zhiyuan Liu, and Maosong Sun. Optima: Optimizing effectiveness and efficiency for LLM-based multi-agent systems, 2024

  132. [140]

    AutoManual: Constructing instruction manuals by LLM agents via interactive environmental learning, 2024

    Minghao Chen, Yihang Li, Yanting Yang, Shiyu Yu, Binbin Lin, and Xiaofei He. AutoManual: Constructing instruction manuals by LLM agents via interactive environmental learning, 2024

  133. [141]

    Agent-pro: Learning to evolve via policy-level reflection and optimization, 2024

    Wenqi Zhang, Ke Tang, Hai Wu, Mengna Wang, Yongliang Shen, Guiyang Hou, Zeqi Tan, Peng Li, Yueting Zhuang, and Weiming Lu. Agent-pro: Learning to evolve via policy-level reflection and optimization, 2024

  134. [142]

    Symbolic learning enables self-evolving agents, 2024

    Wangchunshu Zhou, Yixin Ou, Shengwei Ding, Long Li, Jialong Wu, Tiannan Wang, Jiamin Chen, Shuai Wang, Xiaohua Xu, Ningyu Zhang, Huajun Chen, and Yuchen Eleanor Jiang. Symbolic learning enables self-evolving agents, 2024

  135. [143]

    Autoguide: Automated generation and selection of context-aware guidelines for large language model agents, 2024

    Yao Fu, Dong-Ki Kim, Jaekyeom Kim, Sungryull Sohn, Lajanugen Logeswaran, Kyunghoon Bae, and Honglak Lee. Autoguide: Automated generation and selection of context-aware guidelines for large language model agents, 2024

  136. [144]

    Pace: Two-timescale self-evolution for small language model agents, 2026

    Chen Ling, Pei Chen, Albert Guan, Jiaming Qu, Shayan Ali Akbar, Madhu Gopinathan, and Erwin Cornejo. Pace: Two-timescale self-evolution for small language model agents, 2026

  137. [145]

    Automated design of agentic systems, 2024

    Shengran Hu, Cong Lu, and Jeff Clune. Automated design of agentic systems, 2024

  138. [146]

    Aflow: Automating agentic workflow generation, 2024

    Jiayi Zhang, Jinyu Xiang, Zhaoyang Yu, Fengwei Teng, Xionghui Chen, Jiaqi Chen, Mingchen Zhuge, Xin Cheng, Sirui Hong, Jinlin Wang, Bingnan Zheng, Bang Liu, Yuyu Luo, and Chenglin Wu. Aflow: Automating agentic workflow generation, 2024

  139. [147]

    AgentSquare: Automatic LLM agent search in modular design space, 2024

    Yu Shang, Yu Li, Keyu Zhao, Likai Ma, Jiahe Liu, Fengli Xu, and Yong Li. AgentSquare: Automatic LLM agent search in modular design space, 2024

  140. [148]

    Evoflow: Evolving diverse agentic workflows on the fly, 2025

    Guibin Zhang, Kaijie Chen, Guancheng Wan, Heng Chang, Hong Cheng, Kun Wang, Shuyue Hu, and Lei Bai. Evoflow: Evolving diverse agentic workflows on the fly, 2025

  141. [149]

    AgentNet: Decentralized evolutionary coordination for LLM-based multi-agent systems, 2025

    Yingxuan Yang, Huacan Chai, Shuai Shao, Yuanyi Song, Siyuan Qi, Renting Rui, and Weinan Zhang. AgentNet: Decentralized evolutionary coordination for LLM-based multi-agent systems, 2025

  142. [150]

    Sew: Self-evolving agentic workflows for automated code generation, 2025

    Siwei Liu, Jinyuan Fang, Han Zhou, Yingxu Wang, and Zaiqiao Meng. Sew: Self-evolving agentic workflows for automated code generation, 2025

  143. [151]

    Sepo: Self-evolving prompt agent for system prompt optimiza- tion, 2026

    Wangcheng Tao, Han Wu, and Weng-Fai Wong. Sepo: Self-evolving prompt agent for system prompt optimiza- tion, 2026

  144. [152]

    Autogenesis: A self-evolving agent protocol, 2026

    Wentao Zhang et al. Autogenesis: A self-evolving agent protocol, 2026

  145. [153]

    Darwin: Dynamic agentically rewriting self-improving network, 2026

    Henry Jiang. Darwin: Dynamic agentically rewriting self-improving network, 2026

  146. [154]

    Darwin godel machine: Open-ended evolution of self-improving agents, 2025

    Jenny Zhang, Shengran Hu, Cong Lu, Robert Lange, and Jeff Clune. Darwin godel machine: Open-ended evolution of self-improving agents, 2025. 24

  147. [155]

    Self-distillation bridges distribution gap in language model fine-tuning

    Zhaorui Yang, Qian Liu, Tianyu Pang, Han Wang, Haozhe Feng, Minfeng Zhu, and Wei Chen. Self-distillation bridges distribution gap in language model fine-tuning. InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL), 2024. URLhttps://ar...

  148. [156]

    COPR: Continual human preference learning via optimal policy regularization

    Han Zhang, Lin Gui, Yu Lei, Yuanzhao Zhai, Yehong Zhang, Zhuo Zhang, Yulan He, Hui Wang, Yue Yu, Kam-Fai Wong, Bin Liang, and Ruifeng Xu. COPR: Continual human preference learning via optimal policy regularization. InFindings of the Association for Computational Linguistics: A...

  149. [157]

    CPPO: Continual learning for reinforcement learning with human feedback

    Han Zhang, Yu Lei, Lin Gui, Min Yang, Yulan He, Hui Wang, and Ruifeng Xu. CPPO: Continual learning for reinforcement learning with human feedback. InInternational Conference on Learning Representations (ICLR), 2024

  150. [158]

    TTRL: Test-time reinforcement learning, 2025

    Yuxin Zuo, Kaiyan Zhang, Li Sheng, Shang Qu, Ganqu Cui, Xuekai Zhu, Haozhan Li, Yuchen Zhang, Xinwei Long, Ermo Hua, Biqing Qi, Youbang Sun, Zhiyuan Ma, Lifan Yuan, Ning Ding, and Bowen Zhou. TTRL: Test-time reinforcement learning, 2025. URLhttps://arxiv.org/abs/2504.16084

  151. [159]

    Denser̸=better: Limits of on-policy self-distillation for continual post-training, 2026

    Meng Wang, Haohan Zhao, Wenzhuo Liu, Lu Yang, Geng Liu, Haiyang Guo, Guo-Sen Xie, Gaofeng Meng, Hongbin Liu, and Fei Zhu. Denser̸=better: Limits of on-policy self-distillation for continual post-training, 2026

  152. [160]

    Learning on the job: Test-time curricula for targeted reinforcement learning.arXiv preprint arXiv:2510.04786, 2025

    Jonas Hübotter, Leander Diaz-Bone, Ido Hakimi, Andreas Krause, and Moritz Hardt. Learning on the job: Test-time curricula for targeted reinforcement learning.arXiv preprint arXiv:2510.04786, 2025

  153. [161]

    Self-adapting language models

    Adam Zweiger, Jyothish Pari, Han Guo, Ekin Akyürek, Yoon Kim, and Pulkit Agrawal. Self-adapting language models. InAdvances in Neural Information Processing Systems (NeurIPS), 2025

  154. [162]

    AIMMerging: Adaptive iterative model merging using training trajectories for language model continual learning

    Yujie Feng, Jian Li, Xiaoyu Dong, Pengfei Xu, Xiaohui Zhou, Yujia Zhang, Zexin Lu, Yasha Wang, Alan Zhao, Xu Chu, and Xiao-Ming Wu. AIMMerging: Adaptive iterative model merging using training trajectories for language model continual learning. InProceedings of the 2025 Confere...

  155. [163]

    Mitigating catastrophic forgetting in language transfer via model merging

    Anton Alexandrov, Veselin Raychev, Mark Niklas Müller, Ce Zhang, Martin Vechev, and Kristina Toutanova. Mitigating catastrophic forgetting in language transfer via model merging. InFindings of the Association for Computational Linguistics: EMNLP 2024, 2024

  156. [164]

    Geometry conflict: Explaining and controlling forgetting in LLM continual post-training, 2026

    Yuanyi Wang, Yifan Yang, Su Lu, Yanggan Gu, Pengkai Wang, Wenjun Wang, Zhaoyi Yan, Congkai Xie, Jianmin Wu, Jialun Cao, Shing-Chi Cheung, and Hongxia Yang. Geometry conflict: Explaining and controlling forgetting in LLM continual post-training, 2026

  157. [165]

    ZeroFlow: Overcoming catastrophic forgetting is easier than you think

    Tao Feng, Wei Li, Didi Zhu, Hangjie Yuan, Wendi Zheng, Dan Zhang, and Jie Tang. ZeroFlow: Overcoming catastrophic forgetting is easier than you think. InProceedings of the 42nd International Conference on Machine Learning, volume 267 ofProceedings of Machine Learning Research,...

  158. [166]

    Merge before forget: A single LoRA continual learning via continual merging,

    Fuli Qiao and Mehrdad Mahdavi. Merge before forget: A single LoRA continual learning via continual merging,

  159. [167]

    Evolu- tionary strategies lead to catastrophic forgetting in LLMs, 2026

    Immanuel Abdi, Akshat Gupta, Micah Mok, Alexander Lu, Nicholas Lee, and Gopala Anumanchipalli. Evolu- tionary strategies lead to catastrophic forgetting in LLMs, 2026

  160. [168]

    Suchin Gururangan, Ana Marasović, Swabha Swayamdipta, Kyle Lo, Iz Beltagy, Doug Downey, and Noah A. Smith. Don’t stop pretraining: Adapt language models to domains and tasks. InProceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 2020

  161. [169]

    In- vestigating continual pretraining in large language models: Insights and implications, 2024

    Çağatay Yıldız, Nishaanth Kanna Ravichandran, Prishruit Punia, Matthias Bethge, and Beyza Ermis. In- vestigating continual pretraining in large language models: Insights and implications, 2024. URLhttps: //arxiv.org/abs/2402.17400

  162. [170]

    Method label in Figure 2: SLAO

  163. [171]

    TRACE: A comprehensive benchmark for continual learning in large language models, 2023

    Xiao Wang, Yuansen Zhang, Tianze Chen, Songyang Gao, Senjie Jin, Xianjun Yang, Zhiheng Xi, Rui Zheng, Yicheng Zou, Tao Gui, Qi Zhang, and Xuanjing Huang. TRACE: A comprehensive benchmark for continual learning in large language models, 2023. URLhttps://arxiv.org/abs/2310.06762. 25

  164. [172]

    Spurious forgetting in continual learning of language models, 2025

    Junhao Zheng, Xidi Cai, Shengjie Qiu, and Qianli Ma. Spurious forgetting in continual learning of language models, 2025. URLhttps://arxiv.org/abs/2501.13453

  165. [173]

    Continual gui agents

    Ziwei Liu, Borui Kang, Hangjie Yuan, Zixiang Zhao, Wei Li, Yifan Zhu, and Tao Feng. Continual gui agents. ICML, 2026

  166. [174]

    An empirical study of catastrophic forgetting in large language models during continual fine-tuning.ArXiv, abs/2308.08747, 2023

    Yun Luo, Zhen Yang, Fandong Meng, Yafu Li, Jie Zhou, and Yuechen Zhang. An empirical study of catastrophic forgetting in large language models during continual fine-tuning.ArXiv, abs/2308.08747, 2023. URLhttps: //api.semanticscholar.org/CorpusID:261031244

  167. [175]

    Harness handbook: Making evolving agent harnesses readable, navigable, and editable

    Ruhan Wang, Yucheng Shi, Zongxia Li, Zhongzhi Li, Yue Yu, Junyao Yang, Kishan Panaganti, Haitao Mi, Dongruo Zhou, et al. Harness handbook: Making evolving agent harnesses readable, navigable, and editable. arXiv preprint arXiv:2607.13285, 2026

  168. [176]

    Editing models with task arithmetic.arXiv preprint arXiv:2212.04089, 2022

    Gabriel Ilharco, Marco Tulio Ribeiro, Mitchell Wortsman, Suchin Gururangan, Ludwig Schmidt, Hannaneh Hajishirzi, and Ali Farhadi. Editing models with task arithmetic.arXiv preprint arXiv:2212.04089, 2022

  169. [177]

    TIES-merging: Resolving interference when merging models

    Prateek Yadav, Derek Tam, Leshem Choshen, Colin Raffel, and Mohit Bansal. TIES-merging: Resolving interference when merging models. InThirty-seventh Conference on Neural Information Processing Systems,

  170. [178]

    LoRA: Low-rank adaptation of large language models.ICLR, 1(2):3, 2022

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, et al. LoRA: Low-rank adaptation of large language models.ICLR, 1(2):3, 2022

  171. [179]

    Language models are super mario: Absorbing abilities from homologous models as a free lunch

    Le Yu, Bowen Yu, Haiyang Yu, Fei Huang, and Yongbin Li. Language models are super mario: Absorbing abilities from homologous models as a free lunch. InProceedings of the 41st International Conference on Machine Learning, 2024. URLhttps://arxiv.org/abs/2311.03099

  172. [180]

    Twin-merging: Dynamic integration of modular expertise in model merging

    Zhenyi Lu, Chenghao Fan, Wei Wei, Xiaoye Qu, Dangyang Chen, and Yu Cheng. Twin-merging: Dynamic integration of modular expertise in model merging. InAdvances in Neural Information Processing Systems,

  173. [181]

    Why I do not think AGI is right around the corner.https://www.dwarkesh.com/p/ timelines-june-2025, 2025

    Dwarkesh Patel. Why I do not think AGI is right around the corner.https://www.dwarkesh.com/p/ timelines-june-2025, 2025. URLhttps://www.dwarkesh.com/p/timelines-june-2025. Essay, Dwarkesh Podcast; contextual commentary rather than a primary technical source

  174. [182]

    URLhttps://openreview.net/forum?id=xtaX3WyCj1

  175. [183]

    Evolutionary optimization of model merging recipes.arXiv preprint arXiv:2403.13187, 2024

    Takuya Akiba, Makoto Shing, Yujin Tang, Qi Sun, and David Ha. Evolutionary optimization of model merging recipes.arXiv preprint arXiv:2403.13187, 2024

  176. [184]

    Contra dwarkesh on continual learning.https://www.interconnects.ai/ p/contra-dwarkesh-on-continual-learning, 2025

    Nathan Lambert. Contra dwarkesh on continual learning.https://www.interconnects.ai/ p/contra-dwarkesh-on-continual-learning, 2025. URLhttps://www.interconnects.ai/p/ contra-dwarkesh-on-continual-learning. Essay, Interconnects; contextual commentary rather than a primary techni...

  177. [185]

    CoIN: A benchmark of continual instruction tuning for multimodel large language models

    Cheng Chen, Junchen Zhu, Xu Luo, Heng Tao Shen, Jingkuan Song, and Lianli Gao. CoIN: A benchmark of continual instruction tuning for multimodel large language models. InAdvances in Neural Information Processing Systems, volume 37, pages 57817–57840, 2024. doi: 10.52202/079017-...

  178. [186]

    MLLM-CTBench: A benchmark for continual instruction tuning with reasoning process diagnosis,

    Haiyun Guo, Zhiyan Hou, Yandu Sun, Jinghan He, Yu Chen, Yuzhe Zhou, Yuheng Jia, Jinqiao Wang, and Tat- Seng Chua. MLLM-CTBench: A benchmark for continual instruction tuning with reasoning process diagnosis,

  179. [187]

    Continual test-time domain adaptation

    Qin Wang, Olga Fink, Luc Van Gool, and Dengxin Dai. Continual test-time domain adaptation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022. URLhttps://arxiv.org/ abs/2203.13591. 26 A Glossary of Abbreviations For readability, Table A.1...

  180. [188]

    From storage to experience: A survey on the evolution of LLM agent memory mechanisms, 2026

    Jinghao Luo, Yuchen Tian, Chuxue Cao, Ziyang Luo, Hongzhan Lin, Kaixin Li, Chuyi Kong, Ruichao Yang, and Jing Ma. From storage to experience: A survey on the evolution of LLM agent memory mechanisms, 2026. Findings of ACL 2026

  181. [189]

    The future of continual learning in the era of foundation models: Three key directions,

    Jack Bell, Luigi Quarantiello, Eric Nuertey Coleman, Lanpei Li, Malio Li, Mauro Madeddu, Elia Piccoli, and Vincenzo Lomonaco. The future of continual learning in the era of foundation models: Three key directions,

  182. [190]

    arXiv:2506.03320; accepted at the TCAI workshop, 2025

  183. [194]

    URLhttps://arxiv.org/abs/2508.08275

  184. [2023]

    URLhttps://arxiv.org/abs/2308.04014

  185. [2024]

    URLhttps://arxiv.org/abs/2406.15479

  186. [2025]

    URLhttps://arxiv.org/abs/2509.04259

  187. [2026]

    Method label in Figure 2: SDFT

    URLhttps://arxiv.org/abs/2601.19897. Method label in Figure 2: SDFT

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.