mlsanityData to 7 Oct 2026

Generative AI Act II: Test Time Scaling Drives Cognition Engineering

arXiv 2504.13828 ↗First posted 18 Apr 2025cs.CL cs.AIcited by 14

On the map: LLM reasoning → Reasoning benchmarks and evaluation

Abstract (arXiv)

The first generation of Large Language Models - what might be called "Act I" of generative AI (2020-2023) - achieved remarkable success through massive parameter and data scaling, yet exhibited fundamental limitations such as knowledge latency, shallow reasoning, and constrained cognitive processes. During this era, prompt engineering emerged as our primary interface with AI, enabling dialogue-level communication through natural language. We now witness the emergence of "Act II" (2024-present), where models are transitioning from knowledge-retrieval systems (in latent space) to thought-construction engines through test-time scaling techniques. This new paradigm establishes a mind-level connection with AI through language-based thoughts. In this paper, we clarify the conceptual foundations of cognition engineering and explain why this moment is critical for its development. We systematically break down these advanced approaches through comprehensive tutorials and optimized implementations, democratizing access to cognition engineering and enabling every practitioner to participate in AI's second act. We provide a regularly updated collection of papers on test-time scaling in the GitHub Repository: https://github.com/GAIR-NLP/cognition-engineering

Key facts. Cited by 14 papers in the map (data through 7 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 340 of its 454 references that are in the map. First posted on arXiv 18 Apr 2025.

Builds on

340 of 454 references.

Closest ancestors · 20

Same direction, newest first

  1. 2025A Sober Look at Progress in Language Model Reasoning: Pitfalls and Paths to Reproducibility9 Apr
  2. 2025A Survey on Post-training of Large Language Models8 Mar
  3. 2025An Empirical Study on Eliciting and Improving R1-like Reasoning Models6 Mar
  4. 2025LLM Post-Training: A Deep Dive into Reasoning Large Language Models28 Feb
  5. 2025From System 1 to System 2: A Survey of Reasoning Large Language Models24 Feb
  6. 2025Crowd Comparative Reasoning: Unlocking Comprehensive Evaluations for LLM-as-a-Judge18 Feb
  7. 2025Learning to Plan & Reason for Evaluation with Thinking-LLM-as-a-Judge30 Jan
  8. 2025Reasoning Language Models: A Blueprint20 Jan
  9. 2025Towards Large Reasoning Models: A Survey of Reinforced Reasoning with Large Language Models16 Jan
  10. 2024Can Language Models Learn to Skip Steps?4 Nov
  11. 2024Mutual Reasoning Makes Smaller LLMs Stronger Problem-Solvers12 Aug
  12. 2024Q*: Improving Multi-step Reasoning for LLMs with Deliberative Planning20 Jun
8 more
  1. 2024AutoPSV: Automated Process-Supervised Verifier27 May
  2. 2024LLM Reasoners: New Evaluation, Library, and Analysis of Step-by-Step Reasoning with Large Language Models8 Apr
  3. 2024Quiet-STaR: Language Models Can Teach Themselves to Think Before Speaking14 Mar
  4. 2023Phenomenal Yet Puzzling: Testing Inductive Reasoning Capabilities of Language Models with Hypothesis Refinement12 Oct
  5. 2023At Which Training Stage Does Code Data Help LLMs Reasoning?28 Sep
  6. 2023Hypothesis Search: Inductive Reasoning with Language Models11 Sep
  7. 2023Improving Factuality and Reasoning in Language Models through Multiagent Debate23 May
  8. 2021Show Your Work: Scratchpads for Intermediate Computation with Language Models30 Nov

Other references · 320

  1. 2023GPT-4 Technical Reportcited by 18,132
  2. 2017Proximal Policy Optimization Algorithmscited by 14,441
  3. 2024The Llama 3 Herd of Modelscited by 14,192
  4. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 12,865
  5. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  6. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  7. 2021Evaluating Large Language Models Trained on Codecited by 8,053
  8. 2020Scaling Laws for Neural Language Modelscited by 6,411
  9. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,146
  10. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  11. 2023Mamba: Linear-Time Sequence Modeling with Selective State Spacescited by 4,616
  12. 2023Let's Verify Step by Stepcited by 3,598
38 more
  1. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,748
  2. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,678
  3. 2024LLaVA-OneVision: Easy Visual Task Transfercited by 2,669
  4. 2023A Survey of Large Language Modelscited by 2,537
  5. 2022Do As I Can, Not As I Say: Grounding Language in Robotic Affordancescited by 2,302
  6. 2021Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processingcited by 2,222
  7. 2024HybridFlow: A Flexible and Efficient RLHF Frameworkcited by 1,980
  8. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameterscited by 1,928
  9. 2015High-Dimensional Continuous Control Using Generalized Advantage Estimationcited by 1,916
  10. 2024MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmarkcited by 1,758
  11. 2024A Survey on LLM-as-a-Judgecited by 1,384
  12. 2025Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learningcited by 1,371
  13. 2025s1: Simple test-time scalingcited by 1,311
  14. 2024Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Dualitycited by 1,283
  15. 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,193
  16. 2024Qwen2.5-Coder Technical Reportcited by 1,151
  17. 2024OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environmentscited by 1,134
  18. 2023SGLang: Efficient Execution of Structured Language Model Programscited by 1,107
  19. 2024DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Modelcited by 1,093
  20. 2022Inner Monologue: Embodied Reasoning through Planning with Language Modelscited by 1,057
  21. 2024Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvementcited by 1,057
  22. 2024The AI Scientist: Towards Fully Automated Open-Ended Scientific Discoverycited by 1,002
  23. 2023Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debatecited by 995
  24. 2023Large Language Models Cannot Self-Correct Reasoning Yetcited by 965
  25. 2023Teaching Large Language Models to Self-Debugcited by 963
  26. 2024SimPO: Simple Preference Optimization with a Reference-Free Rewardcited by 951
  27. 2024Large Language Monkeys: Scaling Inference Compute with Repeated Samplingcited by 946
  28. 2024Chameleon: Mixed-Modal Early-Fusion Foundation Modelscited by 904
  29. 2023Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotationscited by 895
  30. 2025Large Language Diffusion Modelscited by 872
  31. 2022Solving math word problems with process- and outcome-based feedbackcited by 871
  32. 2025Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scalingcited by 787
  33. 2024Training Large Language Models to Reason in a Continuous Latent Spacecited by 763
  34. 2023CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquingcited by 751
  35. 2023ChatEval: Towards Better LLM-based Evaluators through Multi-Agent Debatecited by 724
  36. 2023RAFT: Reward rAnked FineTuning for Generative Foundation Model Alignmentcited by 672
  37. 2025SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-trainingcited by 655
  38. 2024Self-Rewarding Language Modelscited by 654

50 most cited of 320 All 320 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 11; 2026: 3.

14 papers.

Sort:

By area:

  1. 2025MegaScience: Pushing the Frontiers of Post-Training Datasets for Science Reasoningcited by 56
  2. 2025Two Experts Are All You Need for Steering Thinking: Reinforcing Cognitive Effort in MoE Reasoning Models Without Additional Trainingcited by 16
  3. 2025OpenDataArena: A Fair and Open Arena for Benchmarking Post-Training Dataset Valuecited by 15
  4. 2025CoRank: LLM-Based Compact Reranking with Document Features for Scientific Retrievalcited by 5
  5. 2025BrowseConf: Confidence-Guided Test-Time Scaling for Web Agentscited by 5
  6. 2025Closing the Data Loop: Using OpenDataArena to Engineer Superior Training Datasetscited by 4
  7. 2025Cog-Rethinker: Hierarchical Metacognitive Reinforcement Learning for LLM Reasoningcited by 3
  8. 2025TrajSelector: Harnessing Latent Representations for Efficient and Effective Best-of-N in Large Reasoning Modelcited by 3
  9. 2026MARTI-MARS2: Scaling Multi-Agent Self-Search via Reinforcement Learning for Code Generationcited by 3
  10. 2025Mitigating Strategy-Selection Bias in Reasoning for More Effective Test-Time Scalingcited by 2
  11. 2025Virus Infection Attack on LLMs: Your Poisoning Can Spread "VIA" Synthetic Datacited by 2
  12. 2026AutoTool: Automatic Scaling of Tool-Use Capabilities in RL via Decoupled Entropy Constraintscited by 2
2 more
  1. 2025Beyond Surface-Level Detection: Towards Cognitive-Driven Defense Against Jailbreak Attacks via Meta-Operations Reasoningcited by 1
  2. 2026Bidirectional Curriculum Generation: A Multi-Agent Framework for Data-Efficient Mathematical Reasoningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026A Taxonomy of Cognitive Capability Gaps in Generative and Agentic AI
  2. 2025The Philosophical Foundations of Growing AI Like A Child
  3. 2023Beyond Words: A Mathematical Framework for Interpreting Large Language Models
  4. 2023Machine Psychology
  5. 2024A Philosophical Introduction to Language Models -- Part I: Continuity With Classic Debates
  6. 2024The Cognitive Capabilities of Generative AI: A Comparative Analysis with Human Benchmarks
  7. 2024CogBench: a large language model walks into a psychology lab
  8. 2025Cognitive AI framework 2.0: advances in the simulation of human thought