A Taxonomy of Cognitive Capability Gaps in Generative and Agentic AI

arXiv 2608.02553 ↗First posted 3 Aug 2026cs.AIcited by 0

On the map: LLM agents and tool use → General LLM agent frameworks and benchmarks

Abstract (arXiv)

Cognitive AI seeks to move beyond language generation and autonomous task execution toward systems capable of sustained reasoning, adaptive behavior, persistent memory, and self-regulation. While generative and agentic AI have demonstrated impressive capabilities across a wide range of tasks, many fundamental cognitive functions remain fragmented or weakly developed, limiting reliable operation over extended time horizons. This paper presents a taxonomy-driven survey of the major cognitive capability gaps that continue to constrain the development of Cognitive AI. The literature is organized around five dimensions: persistent state modeling, goal-directed autonomy, self-monitoring and control, environment interaction, and learning and adaptation. For each dimension, we review recent advances, identify recurring limitations, and discuss open research challenges. Building on these insights, we outline a conceptual Adaptive Cognitive Intelligence Architecture (ACIA) and examine emerging directions in cognition-centric evaluation. The proposed taxonomy provides a unified framework for organizing existing research, identifying unresolved challenges, and guiding the design of future cognitively capable systems. Together, the taxonomy, architectural perspective, and evaluation framework offer a roadmap for advancing AI systems that exhibit more reliable long-term reasoning, adaptive decision-making, and continual learning. The survey highlights key research opportunities toward more adaptive, reliable, and cognitively capable AI systems, providing a foundation for future progress toward Cognitive AI and, ultimately, Artificial General Intelligence (AGI).

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 78 of its 104 references that are in the map. First posted on arXiv 3 Aug 2026.

Builds on

78 of 104 references.

Closest ancestors · 5

Same direction, newest first

Other references · 73

2022Training language models to follow instructions with human feedbackcited by 15,712
2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,393
2016Overcoming catastrophic forgetting in neural networkscited by 4,718
2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,064
2021On the Opportunities and Risks of Foundation Modelscited by 3,961
2023Generative Agents: Interactive Simulacra of Human Behaviorcited by 3,500
2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,317
2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,155
2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,919
2022Locating and Editing Factual Associations in GPTcited by 2,609
2023Sparks of Artificial General Intelligence: Early experiments with GPT-4cited by 2,579

38 more

2022Survey of Hallucination in Natural Language Generationcited by 2,056
2022Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language modelscited by 1,970
2019Transformer-XL: Attentive Language Models Beyond a Fixed-Length Contextcited by 1,830
2023MemGPT: Towards LLMs as Operating Systemscited by 1,295
2022Red Teaming Language Models with Language Modelscited by 983
2022Mass-Editing Memory in a Transformercited by 976
2023Large Language Models Cannot Self-Correct Reasoning Yetcited by 965
2023Graph of Thoughts: Solving Elaborate Problems with Large Language Modelscited by 873
2024Training Large Language Models to Reason in a Continuous Latent Spacecited by 763
2019Compressive Transformers for Long-Range Sequence Modellingcited by 699
2023MemoryBank: Enhancing Large Language Models with Long-Term Memorycited by 539
2019SelectiveNet: A Deep Neural Network with an Integrated Reject Optioncited by 333
2025Memory in the Age of AI Agentscited by 285
2022Contrastive Test-Time Adaptationcited by 258
2023Evaluating the Ripple Effects of Knowledge Editing in Language Modelscited by 238
2024When Can LLMs Actually Correct Their Own Mistakes? A Critical Survey of Self-Correction of LLMscited by 230
2024Continual Learning of Large Language Models: A Comprehensive Surveycited by 228
2024AlphaEdit: Null-Space Constrained Knowledge Editing for Language Modelscited by 224
2023Knowledge Editing for Large Language Models: A Surveycited by 189
2023Mitigating the Alignment Tax of RLHFcited by 181
2024Deep Reinforcement Learning for Robotics: A Survey of Real-World Successescited by 153
2024Know Your Limits: A Survey of Abstention in Large Language Modelscited by 116
2023R-Tuning: Instructing Large Language Models to Say `I Don't Know'cited by 106
2024AriGraph: Learning Knowledge Graph World Models with Episodic Memory for LLM Agentscited by 100
2025Uncertainty Quantification and Confidence Calibration in Large Language Models: A Surveycited by 93
2023The Metacognitive Demands and Opportunities of Generative AIcited by 85
2025PlanGenLLMs: A Modern Survey of LLM Planning Capabilitiescited by 82
2023A Survey on Uncertainty Quantification Methods for Deep Learningcited by 78
2025Reasoning Beyond Language: A Comprehensive Survey on Latent Chain-of-Thought Reasoningcited by 72
2023Think-in-Memory: Recalling and Post-thinking Enable LLMs with Long-Term Memorycited by 64
2024Analyzing and Reducing Catastrophic Forgetting in Parameter Efficient Tuningcited by 59
2023Evaluating and Improving Tool-Augmented Computation-Intensive Math Reasoningcited by 49
2023ChatCoT: Tool-Augmented Chain-of-Thought Reasoning on Chat-based Large Language Modelscited by 43
2023Unveiling the Pitfalls of Knowledge Editing for Large Language Modelscited by 43
2024Learning to Edit: Aligning LLMs with Knowledge Editingcited by 43
2025Building Self-Evolving Agents via Experience-Driven Lifelong Learning: A Framework and Benchmarkcited by 43
2024Envisioning Outlier Exposure by Large Language Models for Out-of-Distribution Detectioncited by 30
2020On the Curse of Memory in Recurrent Neural Networks: Approximation and Optimization Analysiscited by 26

50 most cited of 73 All 73 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way