mlsanityData to 7 Oct 2026

textttSEM-CTRL: Semantically Controlled Decoding

arXiv 2503.01804 ↗First posted 3 Mar 2025cs.CL cs.AI cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Ensuring both syntactic and semantic correctness in Large Language Model (LLM) outputs remains a significant challenge, despite being critical for real-world deployment. In this paper, we introduce textttSEM-CTRL, a unified approach that allows for enforcing rich context-sensitive constraints, and task and instance specific semantics directly on the LLM decoder. Our approach integrates token-level MCTS which is guided by specific syntactic and semantic constraints. The constraints over desired outputs are expressed using Answer Set Grammars, which is a logic-based formalism that generalizes context sensitive grammars while incorporating background knowledge to represent task-specific semantics. We show that our approach helps guarantee valid completions for any off-the-shelf LLM without the need for fine-tuning. We evaluate textttSEM-CTRL on a range of tasks, including synthetic grammar synthesis, combinatorial reasoning, JSON parsing, and planning. Our experimental results demonstrate that textttSEM-CTRL allows even small pre-trained LLMs to efficiently outperform larger variants and state-of-the-art reasoning models (e.g., textito4-mini) while simultaneously guaranteeing semantic validity.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 42 of its 67 references that are in the map. First posted on arXiv 3 Mar 2025.

Builds on

42 of 67 references.

References · 42

  1. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
  3. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,155
  4. 2023Reasoning with Language Model is Planning with World Modelcited by 846
  5. 2017Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithmcited by 723
  6. 2023Language Agent Tree Search Unifies Reasoning Acting and Planning in Language Modelscited by 551
  7. 2021PICARD: Parsing Incrementally for Constrained Auto-Regressive Decoding from Language Modelscited by 402
  8. 2021FUDGE: Controlled Text Generation With Future Discriminatorscited by 374
  9. 2022PlanBench: An Extensible Benchmark for Evaluating Large Language Models on Planning and Reasoning about Changecited by 372
  10. 2023On the Planning Abilities of Large Language Models : A Critical Investigationcited by 308
  11. 2023Tool Learning with Foundation Modelscited by 302
  12. 2023Alphazero-like Tree-Search can Guide Large Language Model Decoding and Trainingcited by 291
30 more
  1. 2017Lexically Constrained Decoding for Sequence Generation Using Grid Beam Searchcited by 234
  2. 2024Spider 2.0: Evaluating Language Models on Real-World Enterprise Text-to-SQL Workflowscited by 223
  3. 2023Large Language Model Guided Tree-of-Thoughtcited by 218
  4. 2023ART: Automatic multi-step reasoning and tool-use for large language modelscited by 193
  5. 2023Grammar-Constrained Decoding for Structured NLP Tasks without Finetuningcited by 186
  6. 2023Planning with Large Language Models for Code Generationcited by 179
  7. 2021Constrained Language Models Yield Few-Shot Semantic Parserscited by 162
  8. 2020MTOP: A Comprehensive Multilingual Task-Oriented Semantic Parsing Benchmarkcited by 156
  9. 2022Synchromesh: Reliable code generation from pre-trained language modelscited by 156
  10. 2021NeuroLogic A*esque Decoding: Constrained Text Generation with Lookahead Heuristicscited by 153
  11. 2024On the Self-Verification Limitations of Large Language Models on Reasoning and Planning Taskscited by 137
  12. 2022Diverse Demonstrations Improve In-context Compositional Generalizationcited by 136
  13. 2024From Decoding to Meta-Generation: Inference-time Algorithms for Large Language Modelscited by 136
  14. 2022Prompting Is Programming: A Query Language for Large Language Modelscited by 110
  15. 2016Guided Open Vocabulary Image Captioning with Constrained Beam Searchcited by 105
  16. 2024Guiding LLMs The Right Way: Fast, Non-Invasive Constrained Generationcited by 99
  17. 2022Compositional Semantic Parsing with Large Language Modelscited by 95
  18. 2023Sequential Monte Carlo Steering of Large Language Models using Probabilistic Programscited by 79
  19. 2023COLLIE: Systematic Construction of Constrained Text Generation Taskscited by 69
  20. 2023Tractable Control for Autoregressive Language Generationcited by 65
  21. 2023Grammar Prompting for Domain-Specific Language Generation with Large Language Modelscited by 58
  22. 2024Grammar-Aligned Decodingcited by 54
  23. 2024Automata-based constraints for language model decodingcited by 49
  24. 2023Symbol-LLM: Towards Foundational Symbol-centric Interface For Large Language Modelscited by 38
  25. 2023Leveraging Large Language Models to Generate Answer Set Programscited by 33
  26. 2024IterGen: Iterative Semantic-aware Structured LLM Generation with Backtrackingcited by 28
  27. 2021The Power of Prompt Tuning for Low-Resource Semantic Parsingcited by 21
  28. 2023REX: Rapid Exploration and eXploitation for AI Agentscited by 13
  29. 2022BenchCLAMP: A Benchmark for Evaluating Language Models on Syntactic and Semantic Parsingcited by 12
  30. 2024Attribute Controlled Fine-tuning for Large Language Models: A Case Study on Detoxificationcited by 7

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Chopthin-Consensus Power Sampling: A Diversity-Preserving Approach to LLM Decoding
  2. 2025Sample Smart, Not Hard: Correctness-First Decoding for Better Reasoning in LLMs
  3. 2026Think Twice Before You Write -- an Entropy-based Decoding Strategy to Enhance LLM Reasoning
  4. 2024SDSAT: Accelerating LLM Inference through Speculative Decoding with Semantic Adaptive Tokens
  5. 2026Thinking Before Constraining: A Unified Decoding Framework for Large Language Models
  6. 2023Don't Fine-Tune, Decode: Syntax Error-Free Tool Use via Constrained Decoding
  7. 2026Constrained Decoding Eliminates Structural Failures in Small LLMs but Reveals a Scale-Dependent Semantic Gap
  8. 2024Adaptable Logical Control for Large Language Models