mlsanityData to 14 Oct 2026

CBF-LLM: Safe Control for LLM Alignment

arXiv 2408.15625 ↗First posted 28 Aug 2024eess.SY cs.AI cs.CLcited by 9

On the map: LLM alignment / RLHF / preference optimisation → LLM preference optimization methods

Abstract (arXiv)

This paper proposes a control-based framework for aligning large language models (LLMs) by leveraging a control barrier function (CBF) to ensure user-desirable text generation. The presented framework applies the safety filter, designed based on the CBF, to the output generation of the baseline LLM, i.e., the sequence of the token, with the aim of intervening in the generated text. The overall text-generation system is implemented with Llama 3 and a RoBERTa model, and the source code is available at https://github.com/Mya-Mya/CBF-LLM. The experiment demonstrates its control ability and effectiveness in reducing the number of interventions needed for user-specified alignment tasks.

Key facts. Cited by 9 papers in the map (data through 14 Oct 2026). Placed in LLM preference optimization methods, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 38 of its 63 references that are in the map. First posted on arXiv 28 Aug 2024.

Builds on

38 of 63 references.

Closest ancestors · 6

Same direction, newest first

  1. 2024DeAL: Decoding-time Alignment for Large Language Models5 Feb
  2. 2023Large Language Model Alignment: A Survey26 Sep
  3. 2023Aligning Large Language Models through Synthetic Feedback23 May
  4. 2023LIMA: Less Is More for Alignment18 May
  5. 2023Aligning Language Models with Preferences through f-divergence Minimization16 Feb
  6. 2022Training language models to follow instructions with human feedback4 Mar

Other references · 32

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2024The Llama 3 Herd of Modelscited by 14,715
  3. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  4. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
  5. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 3,050
  6. 2023G-Eval: NLG Evaluation using GPT-4 with Better Human Alignmentcited by 2,276
  7. 2023Inference-Time Intervention: Eliciting Truthful Answers from a Language Modelcited by 1,302
  8. 2019CTRL: A Conditional Transformer Language Model for Controllable Generationcited by 762
  9. 2023Safe RLHF: Safe Reinforcement Learning from Human Feedbackcited by 656
  10. 2022Improving alignment of dialogue agents via targeted human judgementscited by 531
  11. 2024Large Language Models: A Surveycited by 525
  12. 2021FUDGE: Controlled Text Generation With Future Discriminatorscited by 376
20 more
  1. 2019Control Barrier Functions: Theory and Applicationscited by 372
  2. 2020Autoregressive Entity Retrievalcited by 334
  3. 2019End-to-End Safe Reinforcement Learning through Barrier Functions for Safety-Critical Continuous Control Taskscited by 289
  4. 2024SafeDecoding: Defending against Jailbreak Attacks via Safety-Aware Decodingcited by 247
  5. 2022Safe Control with Learned Certificates: A Survey of Neural Lyapunov, Barrier, and Contraction methodscited by 151
  6. 2023Chain of Hindsight Aligns Language Models with Feedbackcited by 148
  7. 2023Controlled Decoding from Language Modelscited by 135
  8. 2022TimeLMs: Diachronic Language Models from Twittercited by 131
  9. 2020Safety-Critical Model Predictive Control with Discrete-Time Control Barrier Functioncited by 49
  10. 2023What's the Magic Word? A Control Theory of LLM Promptingcited by 41
  11. 2019Adaptive Safety with Control Barrier Functionscited by 25
  12. 2023Taming AI Bots: Controllability of Neural States in Large Language Modelscited by 22
  13. 2024Is In-Context Learning Sufficient for Instruction Following in LLMs?cited by 19
  14. 2024Unlearning Traces the Influential Training Data of Language Modelscited by 18
  15. 2022Control Barrier Functions and Input-to-State Safety with Application to Automated Vehiclescited by 13
  16. 2019Control Barrier Functions for Mechanical Systems: Theory and Application to Robotic Graspingcited by 8
  17. 2024Safety-Critical Control for Autonomous Systems: Control Barrier Functions via Reduced-Order Modelscited by 8
  18. 2022Control Barrier Functions for Stochastic Systems and Safety-critical Control Designscited by 4
  19. 2024Language Model Sentence Completion with a Parser-Driven Rhetorical Control Methodcited by 3
  20. 2022Distributed Coverage Hole Prevention for Visual Environmental Monitoring with Quadcopters via Nonsmooth Control Barrier Functionscited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 1; 2025: 4; 2026: 4.

9 papers.

  1. 2025Steering Dialogue Dynamics for Robustness against Multi-turn Jailbreaking Attackscited by 20
  2. 2025Preemptive Detection and Steering of LLM Misalignment via Latent Reachabilitycited by 10
  3. 2026Local Linearity of LLMs Enables Activation Steering via Model-Based Linear Optimal Controlcited by 9
  4. 2024Generative AI Agents in Autonomous Machines: A Safety Perspectivecited by 7
  5. 2026Activation Steering of Video Generation Models via Reduced-Order Linear Optimal Controlcited by 3
  6. 2025ISACL: Internal State Analyzer for Copyrighted Training Data Leakagecited by 2
  7. 2025From Refusal to Recovery: A Control-Theoretic Approach to Generative AI Guardrailscited by 1
  8. 2026BarrierSteer: LLM Safety via Learning Barrier Steeringcited by 0
  9. 2026Deep learning from a control vantage pointcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Control Barrier Function for Aligning Large Language Models
  2. 2025A Comprehensive Evaluation framework of Alignment Techniques for LLMs
  3. 2026DNAlign: Dynamic Null-Space Safe Alignment for LLMs
  4. 2024Towards Comprehensive Post Safety Alignment of Large Language Models via Safety Patching
  5. 2024Foundational Challenges in Assuring Alignment and Safety of Large Language Models
  6. 2024MULTIVERSE: Exposing Large Language Model Alignment Problems in Diverse Worlds
  7. 2025Where to Start Alignment? Diffusion Large Language Model May Demand a Distinct Position
  8. 2024Towards Scalable Automated Alignment of LLMs: A Survey