Fetching the paper…

Reading the bibliography…

2025

ArcMemo: Abstract Reasoning Composition with Lifelong LLM Memory

Ho, Matthew, Si, Chen, Feng, Zhaoxiang et al.

Original
PDF
Post on X

Understand

While inference-time scaling enables LLMs to carry out increasingly long and capable reasoning traces, the patterns and insights uncovered during these traces are immediately discarded once the context window is reset for a new query.

External memory is a natural way to persist these discoveries, and recent work has shown clear benefits for reasoning-intensive tasks.
We see an opportunity to make such memories more broadly reusable and scalable by moving beyond instance-based memory entries (e.g.
exact query/response pairs, or summaries tightly coupled with the original problem context) toward concept-level memory: reusable, modular abstractions distilled from solution traces and stored in natural language.

ArcMemo: Abstract Reasoning Composition with Lifelong LLM Memory · Around

Built on

Ernie: Enhanced language representation with informative entities, 2019
Original
Zhengyan Zhang, Xu Han, Zhiyuan Liu, Xin Jiang, Maosong Sun, and Qun Liu · 1905
Earlier work this paper cites.
On the measure of intelligence, 2019
Original
François Chollet · 1911
Earlier work this paper cites.
Large scale online learning
Léon Bottou and Yann LeCun · 2003
Earlier work this paper cites.
On-line learning for very large data sets
Léon Bottou and Yann LeCun · 2005
Earlier work this paper cites.
Retrieval-augmented generation for knowledge-intensive nlp tasks, 2021
Original
Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela · 2005
Earlier work this paper cites.
Recurrent memory transformer, 2022
Original
Aydar Bulatov, Yuri Kuratov, and Mikhail S. Burtsev · 2022
Earlier work this paper cites.
Memorizing transformers, 2022
Original
Yuhuai Wu, Markus N. Rabe, DeLesley Hutchins, and Christian Szegedy · 2022
Earlier work this paper cites.
Self-rag: Learning to retrieve, generate, and critique through self-reflection, 2023
Original
Akari Asai, Zeqiu Wu, Yizhong Wang, Avirup Sil, and Hannaneh Hajishirzi · 2023
Earlier work this paper cites.
Think-in-memory: Recalling and post-thinking enable llms with long-term memory, 2023
Original
Lei Liu, Xiaoyan Yang, Yue Shen, Binbin Hu, Zhiqiang Zhang, Jinjie Gu, and Guannan Zhang · 2023
Earlier work this paper cites.
Self-refine: Iterative refinement with self-feedback, 2023
Original
Aman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan, Luyu Gao, Sarah Wiegreffe, Uri Alon, Nouha Dziri, Shrimai Prabhumoye, Yiming Yang, Shashank Gupta, Bodhisattwa Prasad Majumder, Katherine Hermann, Sean Welleck, Amir Yazdanbakhsh, and Peter Clark · 2023
Earlier work this paper cites.
Generative agents: Interactive simulacra of human behavior, 2023
Original
Joon Sung Park, Joseph C. O’Brien, Carrie J. Cai, Meredith Ringel Morris, Percy Liang, and Michael S. Bernstein · 2023
Earlier work this paper cites.

Similar

Reflexion: Language agents with verbal reinforcement learning, 2023
Original
Noah Shinn, Federico Cassano, Edward Berman, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao · 2023
Cited alongside, same era.
Researchagent: Iterative research idea generation over scientific literature with large language models
Original
Jinheon Baek, Sujay Kumar Jauhar, Silviu Cucerzan, and Sung Ju Hwang · 2024
Cited alongside, same era.
Hiagent: Hierarchical working memory management for solving long-horizon agent tasks with large language model
Original
Mengkang Hu, Tianxing Chen, Qiguang Chen, Yao Mu, Wenqi Shao, and Ping Luo · 2024
Cited alongside, same era.
Large language models cannot self-correct reasoning yet, 2024
Original
Jie Huang, Xinyun Chen, Swaroop Mishra, Huaixiu Steven Zheng, Adams Wei Yu, Xinying Song, and Denny Zhou · 2024
Cited alongside, same era.

Then

Buffer of thoughts: Thought-augmented reasoning with large language models, 2024
Original
Ling Yang, Zhaochen Yu, Tianjun Zhang, Shiyi Cao, Minkai Xu, Wentao Zhang, Joseph E. Gonzalez, and Bin Cui · 2024
Later among the works it cites.
The surprising effectiveness of test-time training for few-shot learning, 2025
Original
Ekin Akyürek, Mehul Damani, Adam Zweiger, Linlu Qiu, Han Guo, Jyothish Pari, Yoon Kim, and Jacob Andreas · 2025
Closest in time.
URL https://arcprize.org/leaderboard
ARC-Prize, 2025 · 2025
Closest in time.
Mem0: Building production-ready ai agents with scalable long-term memory
Original
Prateek Chhikara, Dev Khant, Saket Aryan, Taranjeet Singh, and Deshraj Yadav · 2025
Closest in time.
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 · 2025

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…

When can llms actually correct their own mistakes? a critical survey of self-correction of llms, 2024

Ryo Kamoi, Yusen Zhang, Nan Zhang, Jiawei Han, and Rui Zhang · 2024

Cited alongside, same era.

Combining induction and transduction for abstract reasoning, 2024

Wen-Ding Li, Keya Hu, Carter Larsen, Yuqing Wu, Simon Alford, Caleb Woo, Spencer M. Dunn, Hao Tang, Michelangelo Naim, Dat Nguyen, Wei-Long Zheng, Zenna Tavares, Yewen Pu, and Kevin Ellis · 2024

Cited alongside, same era.

Memllm: Finetuning llms to use an explicit read-write memory

Ali Modarressi, Abdullatif Köksal, Ayyoob Imani, Mohsen Fayyaz, and Hinrich Schütze · 2024

Cited alongside, same era.

Memgpt: Towards llms as operating systems, 2024

Charles Packer, Sarah Wooders, Kevin Lin, Vivian Fang, Shishir G. Patil, Ion Stoica, and Joseph E. Gonzalez · 2024

Cited alongside, same era.

Phenomenal yet puzzling: Testing inductive reasoning capabilities of language models with hypothesis refinement, 2024

Linlu Qiu, Liwei Jiang, Ximing Lu, Melanie Sclar, Valentina Pyatkin, Chandra Bhagavatula, Bailin Wang, Yoon Kim, Yejin Choi, Nouha Dziri, and Xiang Ren · 2024

Cited alongside, same era.

Hypothesis search: Inductive reasoning with language models, 2024

Ruocheng Wang, Eric Zelikman, Gabriel Poesia, Yewen Pu, Nick Haber, and Noah D. Goodman · 2024

Cited alongside, same era.

Make llms better zero-shot reasoners: Structure-orientated autonomous reasoning

Pengfei He, Zitao Li, Yue Xing, Yaling Li, Jiliang Tang, and Bolin Ding

Cited in the paper.

Camelot: Towards large language models with training-free consolidated associative memory, 2024b

Zexue He, Leonid Karlinsky, Donghyun Kim, Julian McAuley, Dmitry Krotov, and Rogerio Feris

Cited in the paper.

Closest in time.

A survey of self-evolving agents: On path to artificial super intelligence, 2025

Huanang Gao, Jiayi Geng, Wenyue Hua, Mengkang Hu, Xinzhe Juan, Hongzhang Liu, Shilong Liu, Jiahao Qiu, Xuan Qi, Yiran Wu, Hongru Wang, Han Xiao, Yuhang Zhou, Shaokun Zhang, Jiayi Zhang, Jinyu Xiang, Yixiong Fang, Qiwen Zhao, Dongrui Liu, Qihan Ren, Cheng Qian, Zhenhailong Wang, Minda Hu, Huazheng Wang, Qingyun Wu, Heng Ji, and Mengdi Wang · 2025

Closest in time.

Contextual experience replay for self-improvement of language agents

Yitao Liu, Chenglei Si, Karthik Narasimhan, and Shunyu Yao · 2025

Closest in time.

Self-improving language models for evolutionary program synthesis: A case study on arc-agi, 2025

Julien Pourcel, Cédric Colas, and Pierre-Yves Oudeyer · 2025

Closest in time.

Dynamic cheatsheet: Test-time learning with adaptive memory, 2025

Mirac Suzgun, Mert Yuksekgonul, Federico Bianchi, Dan Jurafsky, and James Zou · 2025

Closest in time.

Hierarchical reasoning model, 2025

Guan Wang, Jin Li, Yuhao Sun, Xing Chen, Changling Liu, Yue Wu, Meng Lu, Sen Song, and Yasin Abbasi Yadkori · 2025

Closest in time.