mlsanityData to 14 Oct 2026

Factored Agents: Decoupling In-Context Learning and Memorization for Robust Tool Use

arXiv 2503.22931 ↗First posted 29 Mar 2025cs.AIcited by 3

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

Abstract (arXiv)

In this paper, we propose a novel factored agent architecture designed to overcome the limitations of traditional single-agent systems in agentic AI. Our approach decomposes the agent into two specialized components: (1) a large language model (LLM) that serves as a high level planner and in-context learner, which may use dynamically available information in user prompts, (2) a smaller language model which acts as a memorizer of tool format and output. This decoupling addresses prevalent issues in monolithic designs, including malformed, missing, and hallucinated API fields, as well as suboptimal planning in dynamic environments. Empirical evaluations demonstrate that our factored architecture significantly improves planning accuracy and error resilience, while elucidating the inherent trade-off between in-context learning and static memorization. These findings suggest that a factored approach is a promising pathway for developing more robust and adaptable agentic AI systems.

Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 15 of its 26 references that are in the map. First posted on arXiv 29 Mar 2025.

Builds on

15 of 26 references.

References · 15

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  2. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  3. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  4. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,182
  5. 2022Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language modelscited by 1,998
  6. 2023ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIscited by 1,947
  7. 2023Gorilla: Large Language Model Connected with Massive APIscited by 1,332
  8. 2019ZeRO: Memory Optimizations Toward Training Trillion Parameter Modelscited by 1,301
  9. 2024Scaling Synthetic Data Creation with 1,000,000,000 Personascited by 417
  10. 2022Data Distributional Properties Drive Emergent In-Context Learning in Transformerscited by 309
  11. 2016A Network-based End-to-End Trainable Task-oriented Dialogue Systemcited by 252
  12. 2024Many-Shot In-Context Learningcited by 206
3 more
  1. 2023Stay on topic with Classifier-Free Guidancecited by 91
  2. 2019Global-to-local Memory Pointer Networks for Task-Oriented Dialoguecited by 47
  3. 2024Undesirable Memorization in Large Language Models: A Surveycited by 29

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 1.

3 papers.

  1. 2025Teaching Language Models To Gather Information Proactivelycited by 14
  2. 2025AI Compute Architecture and Evolution Trendscited by 1
  3. 2026Small Reasoning Models are Instruction Followers in Function Callingcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026The Path Ahead for Agentic AI: Challenges and Opportunities
  2. 2026ToolAnchor: Anchoring Counterfactual Context to Boost Agentic Tool-use Capability
  3. 2025Adaptation of Agentic AI: A Survey of Post-Training, Memory, and Skills
  4. 2025CoDA: A Context-Decoupled Hierarchical Agent with Reinforcement Learning
  5. 2025Intrinsic Memory Agents: Heterogeneous Multi-Agent LLM Systems through Structured Contextual Memory
  6. 2026Agent Memory Distillation: Empowering Small LLM Agents with Hierarchical Teacher Memory
  7. 2026MAPLE: A Sub-Agent Architecture for Memory, Learning, and Personalization in Agentic AI Systems
  8. 2025Unifying Dynamic Tool Creation and Cross-Task Experience Sharing through Cognitive Memory Architecture