# From Found to Designed: Concepts as a Design Axis for Large Language Models

arXiv: https://arxiv.org/abs/2607.26825 · HTML: https://mlsanity.com/paper/2607.26825/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 33 of its 40 references that are in the map.
- First posted on arXiv 29 Jul 2026.

## Builds on (most cited first, 30 of 33)

- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [Concept Bottleneck Models](https://mlsanity.com/paper/2007.04612/) (2020)
- [Graph of Thoughts: Solving Elaborate Problems with Large Language Models](https://mlsanity.com/paper/2308.09687/) (2023)
- [Probing Classifiers: Promises, Shortcomings, and Advances](https://mlsanity.com/paper/2102.12452/) (2021)
- [ERNIE: Enhanced Language Representation with Informative Entities](https://mlsanity.com/paper/1905.07129/) (2019)
- [Faith and Fate: Limits of Transformers on Compositionality](https://mlsanity.com/paper/2305.18654/) (2023)
- [QA-GNN: Reasoning with Language Models and Knowledge Graphs for Question Answering](https://mlsanity.com/paper/2104.06378/) (2021)
- [StructGPT: A General Framework for Large Language Model to Reason over Structured Data](https://mlsanity.com/paper/2305.09645/) (2023)
- [K-BERT: Enabling Language Representation with Knowledge Graph](https://mlsanity.com/paper/1909.07606/) (2019)
- [Think-on-Graph: Deep and Responsible Reasoning of Large Language Model on Knowledge Graph](https://mlsanity.com/paper/2307.07697/) (2023)
- [Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet](https://mlsanity.com/paper/2605.29358/) (2026)
- [Hierarchical Transformers Are More Efficient Language Models](https://mlsanity.com/paper/2110.13711/) (2021)
- [Large Concept Models: Language Modeling in a Sentence Representation Space](https://mlsanity.com/paper/2412.08821/) (2024)
- [CoLAKE: Contextualized Language and Knowledge Embedding](https://mlsanity.com/paper/2010.00309/) (2020)
- [Verbalizable Representations Form a Global Workspace in Language Models](https://mlsanity.com/paper/2607.15495/) (2026)
- [SenseBERT: Driving Some Sense into BERT](https://mlsanity.com/paper/1908.05646/) (2019)
- [Emotion Concepts and their Function in a Large Language Model](https://mlsanity.com/paper/2604.07729/) (2026)
- [Knowledge-Aware Language Model Pretraining](https://mlsanity.com/paper/2007.00655/) (2020)
- [Concept Bottleneck Large Language Models](https://mlsanity.com/paper/2412.07992/) (2024)
- [RAG-Star: Enhancing Deliberative Reasoning with Retrieval Augmented Verification and Refinement](https://mlsanity.com/paper/2412.12881/) (2024)
- [GraphEval: A Knowledge-Graph Based LLM Hallucination Evaluation Framework](https://mlsanity.com/paper/2407.10793/) (2024)
- [Knowledge Graphs, Large Language Models, and Hallucinations: An NLP Perspective](https://mlsanity.com/paper/2411.14258/) (2024)
- [We Can't Understand AI Using our Existing Vocabulary](https://mlsanity.com/paper/2502.07586/) (2025)
- [Mitigating Hallucinations in Large Language Models via Self-Refinement-Enhanced Knowledge Retrieval](https://mlsanity.com/paper/2405.06545/) (2024)
- [Next Concept Prediction in Discrete Latent Space Leads to Stronger Language Models](https://mlsanity.com/paper/2602.08984/) (2026)
- [FAME: Flexible, Scalable Analogy Mappings Engine](https://mlsanity.com/paper/2311.01860/) (2023)
- [Cooking Up Creativity: Enhancing LLM Creativity through Structured Recombination](https://mlsanity.com/paper/2504.20643/) (2025)
- [Beyond Tokens: Concept-Level Training Objectives for LLMs](https://mlsanity.com/paper/2601.11791/) (2026)
- [When to Think Fast and Slow? AMOR: Adaptive Entropy Gate for Hybrid Models](https://mlsanity.com/paper/2602.13215/) (2026)
- [Improving Large Language Models with Concept-Aware Fine-Tuning](https://mlsanity.com/paper/2506.07833/) (2025)

## Abstract (arXiv)

Large language models (LLMs) encode rich concept-like information, but represent it implicitly through distributed statistical associations rather than as explicit, structured, compositional concepts. Consequently, concept-level structure is typically found rather than designed: it is recovered after training through probing or dictionary learning, with no architectural guarantee of stability, compositionality, controllability, or alignment with human conceptual organization. We organize concept-aware interventions along two dimensions: whether concept structure is internally induced or externally grounded, and the stage of the pipeline where it is introduced. This taxonomy reveals three broad patterns: inference-time approaches remain comparatively underexplored, related ideas have developed largely in isolation across pipeline stages, and externally grounded methods span the entire pipeline despite often being described under different terminology. Together, these observations motivate moving beyond recovering concept-like structure from trained models toward designing LLMs with explicit conceptual representations.