mlsanityData to 7 Oct 2026

Tractable Control for Autoregressive Language Generation

arXiv 2304.07438 ↗First posted 15 Apr 2023cs.CL cs.AIcited by 65

Not on the map. Why?

Abstract (arXiv)

Despite the success of autoregressive large language models in text generation, it remains a major challenge to generate text that satisfies complex constraints: sampling from the conditional distribution Pr(text | α) is intractable for even the simplest lexical constraints α. To overcome this challenge, we propose to use tractable probabilistic models (TPMs) to impose lexical constraints in autoregressive text generation models, which we refer to as GeLaTo (Generating Language with Tractable Constraints). To demonstrate the effectiveness of this framework, we use distilled hidden Markov models, where we can efficiently compute Pr(text | α), to guide autoregressive generation from GPT2. GeLaTo achieves state-of-the-art performance on challenging benchmarks for constrained text generation (e.g., CommonGen), beating various strong baselines by a large margin. Our work not only opens up new avenues for controlling large language models but also motivates the development of more expressive TPMs.

Key facts. Cited by 65 papers in the map (data through 7 Oct 2026). Builds on 31 of its 59 references that are in the map. First posted on arXiv 15 Apr 2023.

Builds on

31 of 59 references.

References · 31

  1. 2020Language Models are Few-Shot Learnerscited by 32,199
  2. 2019BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehensioncited by 5,874
  3. 2014Neural Machine Translation by Jointly Learning to Align and Translatecited by 4,530
  4. 2014CIDEr: Consensus-based Image Description Evaluationcited by 2,027
  5. 2015Effective Approaches to Attention-based Neural Machine Translationcited by 1,024
  6. 2016SPICE: Semantic Propositional Image Caption Evaluationcited by 820
  7. 2021FUDGE: Controlled Text Generation With Future Discriminatorscited by 374
  8. 2018Plan-And-Write: Towards Better Automatic Storytellingcited by 232
  9. 2018Fast Lexically Constrained Decoding with Dynamic Beam Allocation for Neural Machine Translationcited by 181
  10. 2021NeuroLogic A*esque Decoding: Constrained Text Generation with Lookahead Heuristicscited by 153
  11. 2020NeuroLogic Decoding: (Un)supervised Neural Text Generation with Predicate Logic Constraintscited by 146
  12. 2022Semantic Probabilistic Layers for Neuro-Symbolic Learningcited by 103
19 more
  1. 2020Einsum Networks: Fast and Scalable Learning of Tractable Probabilistic Circuitscited by 102
  2. 2022On the Paradox of Learning to Reason from Datacited by 93
  3. 2020Lexically Constrained Neural Machine Translation with Levenshtein Transformercited by 45
  4. 2018Toward Abstractive Summarization Using Semantic Representationscited by 43
  5. 2022Controllable Text Generation with Neurally-Decomposed Oraclecited by 41
  6. 2019Neural Extractive Text Summarization with Syntactic Compressioncited by 36
  7. 2022Scaling Up Probabilistic Circuits by Latent Variable Distillationcited by 35
  8. 2019SPFlow: An Easy and Extensible Library for Deep Probabilistic Learning using Sum-Product Networkscited by 26
  9. 2021Probabilistic Generating Circuitscited by 24
  10. 2022Zero-shot Sonnet Generation with Discourse-level Planning and Aesthetics Featurescited by 24
  11. 2020Group Fairness by Probabilistic Modeling with Latent Fair Decisionscited by 22
  12. 2020Scaling Hidden Markov Language Modelscited by 22
  13. 2020Joints in Random Forestscited by 21
  14. 2019On Tractable Computation of Expected Predictionscited by 19
  15. 2017Boosted Generative Modelscited by 18
  16. 2022Sparse Probabilistic Circuits via Pruning and Growingcited by 17
  17. 2019Learning Fair Naive Bayes Classifiers by Discovering and Eliminating Discrimination Patternscited by 14
  18. 2021InsNet: An Efficient, Flexible, and Performant Insertion-based Text Generation Modelcited by 14
  19. 2022Predictive Querying for Autoregressive Neural Sequence Modelscited by 5

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 10; 2024: 21; 2025: 20; 2026: 14.

65 papers.

Sort:

By area:

  1. 2023LINC: A Neurosymbolic Approach for Logical Reasoning by Combining Language Models with First-Order Logic Proverscited by 161
  2. 2023Sequential Monte Carlo Steering of Large Language Models using Probabilistic Programscited by 79
  3. 2025Theoretical Benefit and Limitation of Diffusion Language Modelcited by 46
  4. 2023Can LLMs Follow Simple Rules?cited by 45
  5. 2025Syntactic and Semantic Control of Large Language Models via Sequential Monte Carlocited by 45
  6. 2023Subtractive Mixture Models via Squaring: Representation and Learningcited by 39
  7. 2024Large Language Models for Medicine: A Surveycited by 38
  8. 2024DeepEdit: Knowledge Editing as Decoding with Constraintscited by 33
  9. 2024Adaptable Logical Control for Large Language Modelscited by 30
  10. 2025Diffusion vs. Autoregressive Language Models: A Text Embedding Perspectivecited by 23
  11. 2023Image Inpainting via Tractable Steering of Diffusion Modelscited by 22
  12. 2024What is the Relationship between Tensor Factorizations and Circuits (and How Can We Exploit it)?cited by 22
38 more
  1. 2023A Pseudo-Semantic Loss for Autoregressive Models with Logical Constraintscited by 21
  2. 2025Fast Controlled Generation from Language Models with Adaptive Weighted Rejection Samplingcited by 20
  3. 2024HYSYNTH: Context-Free LLM Approximation for Guiding Program Synthesiscited by 14
  4. 2024On the Relationship Between Monotone and Squared Probabilistic Circuitscited by 14
  5. 2025Constrained Sampling for Language Models Should Be Easy: An MCMC Perspectivecited by 14
  6. 2023Next Steps for Human-Centered Generative AI: A Technical Perspectivecited by 13
  7. 2024Building Expressive and Tractable Probabilistic Generative Models: A Reviewcited by 12
  8. 2025Neurosymbolic Diffusion Modelscited by 12
  9. 2024Controllable Generation via Locally Constrained Resamplingcited by 11
  10. 2026Breaking the Factorization Barrier in Diffusion Language Modelscited by 11
  11. 2024A Compositional Atlas for Algebraic Circuitscited by 10
  12. 2024CoDa: Constrained Generation based Data Augmentation for Low-Resource NLPcited by 9
  13. 2024Scaling Continuous Latent Variable Models as Probabilistic Integral Circuitscited by 9
  14. 2024NeSyA: Neurosymbolic Automatacited by 9
  15. 2025Scaling Probabilistic Circuits via Monarch Matricescited by 9
  16. 2025Fast and Expressive Multi-Byte Prediction with Probabilistic Circuitscited by 9
  17. 2025TRACE Back from the Future: A Probabilistic Reasoning Approach to Controllable Language Generationcited by 8
  18. 2025Decoupling Task-Solving and Output Formatting in LLM Generationcited by 8
  19. 2026Mitigating Bias in Locally Constrained Decoding via Tractable Proposalscited by 8
  20. 2024Attribute Controlled Fine-tuning for Large Language Models: A Case Study on Detoxificationcited by 7
  21. 2024Guaranteed Generation from Large Language Modelscited by 7
  22. 2024Restructuring Tractable Probabilistic Circuitscited by 7
  23. 2024Control Large Language Models via Divide and Conquercited by 5
  24. 2024Jet Expansions of Residual Computationcited by 4
  25. 2025Tractable Representation Learning with Probabilistic Circuitscited by 4
  26. 2024Approximately Aligned Decodingcited by 3
  27. 2025ABS: Enforcing Constraint Satisfaction On Generated Sequences Via Automata-Guided Beam Searchcited by 3
  28. 2025Tractable Sharpness-Aware Learning of Probabilistic Circuitscited by 3
  29. 2025Constrained Adaptive Rejection Samplingcited by 3
  30. 2024DANA: Domain-Aware Neurosymbolic Agents for Consistency and Accuracycited by 2
  31. 2024Interacting Large Language Model Agents. Interpretable Models and Social Learningcited by 2
  32. 2024On Faster Marginalization with Squared Circuits via Orthonormalizationcited by 2
  33. 2025The Limits of Tractable Marginalizationcited by 2
  34. 2023DiNADO: Norm-Disentangled Neurally-Decomposed Oracles for Controlling Language Modelscited by 1
  35. 2023BOOST: Harnessing Black-Box Control to Boost Commonsense in LMs' Generationcited by 1
  36. 2023Towards Consistent Language Models Using Declarative Constraintscited by 1
  37. 2025Interleaved Gibbs Diffusion: Generating Discrete-Continuous Data with Implicit Constraintscited by 1
  38. 2026Binary Rewards and Reinforcement Learning: Fundamental Challengescited by 1

50 most cited of 65 All 65 (JSON)