mlsanityData to 7 Oct 2026

KARMA: Knowledge graph-based Automated Reasoning Materialization and Alignment

arXiv 2607.03166 ↗First posted 3 Jul 2026cs.CL cs.AI cs.LGcited by 0

On the map: LLM reasoning → Knowledge- and tool-augmented reasoning

Abstract (arXiv)

Template-based contrastive synthesis is scalable, but its candidates often differ only in a few entity-slots while sequence-level optimization spreads supervision over mostly shared templates. We formalize this as the Resolution Mismatch Problem and propose KARMA, which enumerates schema-constrained paths over domain knowledge graphs and verbalizes them into slot-aligned contrastive candidates. Slot-Parallel Alignment (SPA) then applies a decoupled slot-level objective to route preference supervision to discriminative entity-slots, with slot-aware masked attention serving as an optional packed-evaluation implementation. Across biomedical, computer-science, and chemistry benchmarks, KARMA outperforms base LLM and same-data SFT baselines, and compares favorably with sequence- and token-level preference methods.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Knowledge- and tool-augmented reasoning, one of 17 directions in LLM reasoning. Builds on 40 of its 47 references that are in the map. First posted on arXiv 3 Jul 2026.

Builds on

40 of 47 references.

Closest ancestors · 4

Same direction, newest first

  1. 2024Paths-over-Graph: Knowledge Graph Empowered Large Language Model Reasoning18 Oct
  2. 2024Graph Chain-of-Thought: Augmenting Large Language Models by Reasoning on Graphs10 Apr
  3. 2023Reasoning on Graphs: Faithful and Interpretable Large Language Model Reasoning2 Oct
  4. 2023Think-on-Graph: Deep and Responsible Reasoning of Large Language Model on Knowledge Graph15 Jul

Other references · 36

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
  2. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  3. 2020Measuring Massive Multitask Language Understandingcited by 7,768
  4. 2023Let's Verify Step by Stepcited by 3,598
  5. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,748
  6. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,674
  7. 2024MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmarkcited by 1,758
  8. 2020What Disease does this Patient Have? A Large-scale Open Domain Question Answering Dataset from Medical Examscited by 1,324
  9. 2019PubMedQA: A Dataset for Biomedical Research Question Answeringcited by 1,271
  10. 2023WizardLM: Empowering large pre-trained language models to follow complex instructionscited by 1,116
  11. 2023Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotationscited by 895
  12. 2022Solving math word problems with process- and outcome-based feedbackcited by 871
24 more
  1. 2023An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-tuningcited by 622
  2. 2023WizardMath: Empowering Mathematical Reasoning for Large Language Models via Reinforced Evol-Instructcited by 618
  3. 2023Preference Ranking Optimization for Human Alignmentcited by 283
  4. 2023The False Promise of Imitating Proprietary LLMscited by 254
  5. 2024From r to Q^*: Your Language Model is Secretly a Q-Functioncited by 242
  6. 2024Token-level Direct Preference Optimizationcited by 135
  7. 2024A Note on Shumailov et al. (2024): `AI Models Collapse When Trained on Recursively Generated Data'cited by 114
  8. 2024LiPO: Listwise Preference Optimization through Learning-to-Rankcited by 94
  9. 2024Critical Tokens Matter: Token-Level Contrastive Estimation Enhances LLM's Reasoning Capabilitycited by 80
  10. 2023Do Models Really Learn to Follow Instructions? An Empirical Study of Instruction Tuningcited by 48
  11. 2025The Best Instruction-Tuning Data are Those That Fitcited by 48
  12. 2024TIS-DPO: Token-level Importance Sampling for Direct Preference Optimization With Estimated Weightscited by 42
  13. 2025Scaling Laws for Forgetting during Finetuning with Pretraining Data Injectioncited by 35
  14. 2024CS-Bench: A Comprehensive Benchmark for Large Language Models towards Computer Science Masterycited by 30
  15. 2024Selective Preference Optimization via Token-Level Reward Function Estimationcited by 25
  16. 2024Learning to Plan for Retrieval-Augmented Large Language Models from Knowledge Graphscited by 24
  17. 2025Token-Importance Guided Direct Preference Optimizationcited by 14
  18. 2025Rethinking DPO: The Role of Rejected Responses in Preference Misalignmentcited by 14
  19. 2024T-REG: Preference Optimization with Token-Level Reward Regularizationcited by 13
  20. 2024OCEAN: Offline Chain-of-thought Evaluation and Alignment in Large Language Modelscited by 12
  21. 2024GLaM: Fine-Tuning Large Language Models for Domain Knowledge Graph Alignment via Neighborhood Partitioning and Generative Subgraph Encodingcited by 11
  22. 2024SparsePO: Controlling Preference Alignment of LLMs via Sparse Token Maskscited by 9
  23. 2024Generating Uncontextualized and Contextualized Questions for Document-Level Event Argument Extractioncited by 5
  24. 2025TGDPO: Harnessing Token-Level Reward Guidance for Enhancing Direct Preference Optimizationcited by 5

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Reason-Align-Respond: Aligning LLM Reasoning with Knowledge Graphs for KGQA
  2. 2025In-situ graph reasoning and knowledge expansion using Graph-PReFLexOR
  3. 2025Hydra: A Modular Architecture for Efficient Long-Context Reasoning
  4. 2026Query-Conditioned Knowledge Alignment for Reliable Cross-System Medical Reasoning
  5. 2026TDA-RC: Task-Driven Alignment for Knowledge-Based Reasoning Chains in Large Language Models
  6. 2025MedRule-KG: A Knowledge-Graph--Steered Scaffold for Reliable Mathematical and Biomedical Reasoning
  7. 2026GRACE: Gradient-aligned Reasoning Data Curation for Efficient Post-training
  8. 2025Enhancing Large Language Model for Knowledge Graph Completion via Structure-Aware Alignment-Tuning