mlsanityData to 7 Oct 2026

Learning to Reason and Memorize with Self-Notes

arXiv 2305.00833 ↗First posted 1 May 2023cs.LG cs.AI cs.CLcited by 30

On the map: LLM reasoning → Reasoning benchmarks and evaluation

Abstract (arXiv)

Large language models have been shown to struggle with multi-step reasoning, and do not retain previous reasoning steps for future use. We propose a simple method for solving both of these problems by allowing the model to take Self-Notes. Unlike recent chain-of-thought or scratchpad approaches, the model can deviate from the input context at any time to explicitly think and write down its thoughts. This allows the model to perform reasoning on the fly as it reads the context and even integrate previous reasoning steps, thus enhancing its memory with useful information and enabling multi-step reasoning. Experiments across a wide variety of tasks demonstrate that our method can outperform chain-of-thought and scratchpad methods by taking Self-Notes that interleave the input text.

Key facts. Cited by 30 papers in the map (data through 7 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 39 of its 50 references that are in the map. First posted on arXiv 1 May 2023.

Builds on

39 of 50 references.

Closest ancestors · 3

Same direction, newest first

  1. 2022Measuring and Narrowing the Compositionality Gap in Language Models7 Oct
  2. 2022Selection-Inference: Exploiting Large Language Models for Interpretable Logical Reasoning19 May
  3. 2021Show Your Work: Scratchpads for Intermediate Computation with Language Models30 Nov

Other references · 36

  1. 2017Attention Is All You Needcited by 59,932
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2020Language Models are Few-Shot Learnerscited by 32,199
  4. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
  5. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  6. 2022PaLM: Scaling Language Modeling with Pathwayscited by 5,304
  7. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,581
  8. 2022Large Language Models are Zero-Shot Reasonerscited by 4,323
  9. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,072
  10. 2021Program Synthesis with Large Language Modelscited by 3,509
  11. 2023Sparks of Artificial General Intelligence: Early experiments with GPT-4cited by 2,579
  12. 2022Do As I Can, Not As I Say: Grounding Language in Robotic Affordancescited by 2,302
24 more
  1. 2021Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolationcited by 1,085
  2. 2022Inner Monologue: Embodied Reasoning through Planning with Language Modelscited by 1,057
  3. 2022Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questionscited by 911
  4. 2022STaR: Bootstrapping Reasoning With Reasoningcited by 895
  5. 2017Program Induction by Rationale Generation : Learning to Solve and Explain Algebraic Word Problemscited by 642
  6. 2018Universal Transformerscited by 613
  7. 2016Adaptive Computation Time for Recurrent Neural Networkscited by 479
  8. 2018e-SNLI: Natural Language Inference with Natural Language Explanationscited by 468
  9. 2016Solving General Arithmetic Word Problemscited by 353
  10. 2015Towards AI-Complete Question Answering: A Set of Prerequisite Toy Taskscited by 309
  11. 2019Compositionality decomposed: how do neural networks generalise?cited by 264
  12. 2022Exploring Length Generalization in Large Language Modelscited by 204
  13. 2019Levenshtein Transformercited by 197
  14. 2017Adaptive Neural Networks for Efficient Inferencecited by 161
  15. 2021PonderNet: Learning to Pondercited by 156
  16. 2019Insertion Transformer: Flexible Sequence Generation via Insertion Operationscited by 123
  17. 2022Block-Recurrent Transformerscited by 107
  18. 2022PEER: A Collaborative Language Modelcited by 86
  19. 2021Chess as a Testbed for Language Model State Trackingcited by 69
  20. 2016Tracking the World State with Recurrent Entity Networkscited by 51
  21. 2021SHAPE: Shifted Absolute Position Embedding for Transformerscited by 37
  22. 2022Generating Information-Seeking Conversations from Unlabeled Documentscited by 17
  23. 2022The Curious Case of Absolute Position Embeddingscited by 17
  24. 2021Staircase Attention for Recurrent Processing of Sequencescited by 13

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 5; 2024: 16; 2025: 6; 2026: 3.

30 papers.

Sort:

By area:

  1. 2023Chain-of-Verification Reduces Hallucination in Large Language Modelscited by 317
  2. 2024Quiet-STaR: Language Models Can Teach Themselves to Think Before Speakingcited by 292
  3. 2023Think before you speak: Training Language Models With Pause Tokenscited by 254
  4. 2024Evaluation of Retrieval-Augmented Generation: A Surveycited by 146
  5. 2024A Human-Inspired Reading Agent with Gist Memory of Very Long Contextscited by 140
  6. 2023Walking Down the Memory Maze: Beyond Context Limit through Interactive Readingcited by 129
  7. 2024HiAgent: Hierarchical Working Memory Management for Solving Long-Horizon Agent Tasks with Large Language Modelcited by 101
  8. 2023Exposing Attention Glitches with Flip-Flop Language Modelingcited by 74
  9. 2024Self-Taught Evaluatorscited by 70
  10. 2024Thinking LLMs: General Instruction Following with Thought Generationcited by 45
  11. 2024How Far Can Transformers Reason? The Globality Barrier and Inductive Scratchpadcited by 35
  12. 2025LLM Pretraining with Continuous Conceptscited by 33
18 more
  1. 2024Surveying the Effects of Quality, Diversity, and Complexity in Synthetic Data From Large Language Modelscited by 31
  2. 2024Universal Length Generalization with Turing Programscited by 25
  3. 2024TOOLVERIFIER: Generalization to New Tools via Self-Verificationcited by 23
  4. 2025Retrieval Augmented Generation Evaluation in the Era of Large Language Models: A Comprehensive Surveycited by 20
  5. 2023Dissecting Chain-of-Thought: Compositionality through In-Context Filtering and Learningcited by 19
  6. 2024Symbolic Working Memory Enhances Language Models for Complex Rule Applicationcited by 17
  7. 2025RL for Reasoning by Adaptively Revealing Rationalescited by 15
  8. 2024Benchmarking GPT-4 on Algorithmic Problems: A Systematic Evaluation of Prompting Strategiescited by 9
  9. 2024Enhancing Robustness in Large Language Models: Prompting for Mitigating the Impact of Irrelevant Informationcited by 8
  10. 2025The Role of Sparsity for Length Generalization in Transformerscited by 8
  11. 2025On the Failure of Latent State Persistence in Large Language Modelscited by 8
  12. 2024Generating Diverse Hypotheses for Inductive Reasoningcited by 6
  13. 2024LoMA: Lossless Compressed Memory Attentioncited by 5
  14. 2026AutoMem: Automated Learning of Memory as a Cognitive Skillcited by 4
  15. 2024Interacting Large Language Model Agents. Interpretable Models and Social Learningcited by 2
  16. 2025Catch Your Breath: Adaptive Computation for Self-Paced Sequence Productioncited by 2
  17. 2026SCOUT: Active Information Foraging for Long-Text Understanding with Decoupled Epistemic Statescited by 1
  18. 2026AgentSpec: Understanding Embodied Agent Scaffolds Through Controlled Compositioncited by 1

Similar, not cited

Close in wording, no citation either way

  1. 2024Self-Discover: Large Language Models Self-Compose Reasoning Structures
  2. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Models
  3. 2023Self-Polish: Enhance Reasoning in Large Language Models via Problem Refinement
  4. 2025Advancing Reasoning in Large Language Models: Promising Methods and Approaches
  5. 2023MoT: Memory-of-Thought Enables ChatGPT to Self-Improve
  6. 2026Unlocking the Working Memory of Large Language Models for Latent Reasoning
  7. 2024On Memorization of Large Language Models in Logical Reasoning
  8. 2024Self-Harmonized Chain of Thought