mlsanityData to 7 Oct 2026

A Causal Explainable Guardrails for Large Language Models

arXiv 2405.04160 ↗First posted 7 May 2024cs.CLcited by 18

On the map: LLM safety and security → Safety alignment, refusal and guardrails

Abstract (arXiv)

Large Language Models (LLMs) have shown impressive performance in natural language tasks, but their outputs can exhibit undesirable attributes or biases. Existing methods for steering LLMs toward desired attributes often assume unbiased representations and rely solely on steering prompts. However, the representations learned from pre-training can introduce semantic biases that influence the steering process, leading to suboptimal results. We propose LLMGuardrail, a novel framework that incorporates causal analysis and adversarial learning to obtain unbiased steering representations in LLMs. LLMGuardrail systematically identifies and blocks the confounding effects of biases, enabling the extraction of unbiased steering representations. Additionally, it includes an explainable component that provides insights into the alignment between the generated output and the desired direction. Experiments demonstrate LLMGuardrail's effectiveness in steering LLMs toward desired attributes while mitigating biases. Our work contributes to the development of safe and reliable LLMs that align with desired attributes.

Key facts. Cited by 18 papers in the map (data through 7 Oct 2026). Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security. Builds on 60 of its 85 references that are in the map. First posted on arXiv 7 May 2024.

Builds on

60 of 85 references.

Closest ancestors · 1

Same direction, newest first

  1. 2023Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!5 Oct

Other references · 59

  1. 2020Language Models are Few-Shot Learnerscited by 32,199
  2. 2022Training language models to follow instructions with human feedbackcited by 15,712
  3. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 12,865
  4. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  5. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,919
  6. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 2,914
  7. 2022Locating and Editing Factual Associations in GPTcited by 2,609
  8. 2014Unsupervised Domain Adaptation by Backpropagationcited by 1,906
  9. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,582
  10. 2023Jailbreaking Black Box Large Language Models in Twenty Queriescited by 1,326
  11. 2016Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddingscited by 1,304
  12. 2016Understanding intermediate layers using linear classifier probescited by 1,279
38 more
  1. 2023Inference-Time Intervention: Eliciting Truthful Answers from a Language Modelcited by 1,259
  2. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,239
  3. 2019BERT Rediscovers the Classical NLP Pipelinecited by 1,046
  4. 2021Probing Classifiers: Promises, Shortcomings, and Advancescited by 831
  5. 2019Plug and Play Language Models: A Simple Approach to Controlled Text Generationcited by 793
  6. 2022Discovering Latent Knowledge in Language Models Without Supervisioncited by 733
  7. 2023Time-LLM: Time Series Forecasting by Reprogramming Large Language Modelscited by 680
  8. 2023The Internal State of an LLM Knows When It's Lyingcited by 644
  9. 2023Eliciting Latent Predictions from Transformers with the Tuned Lenscited by 579
  10. 2022ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detectioncited by 419
  11. 2021BOLD: Dataset and Metrics for Measuring Biases in Open-Ended Language Generationcited by 410
  12. 2023Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrationscited by 401
  13. 2023DeepInception: Hypnotize Large Language Model to Be Jailbreakercited by 337
  14. 2023Explainability for Large Language Models: A Surveycited by 331
  15. 2024INSIDE: LLMs' Internal States Retain the Power of Hallucination Detectioncited by 317
  16. 2023Language Models Represent Space and Timecited by 306
  17. 2023MQuAKE: Assessing Knowledge Editing in Language Models via Multi-Hop Questionscited by 292
  18. 2019Evaluating Gender Bias in Machine Translationcited by 247
  19. 2022Extracting Latent Steering Vectors from Pretrained Language Modelscited by 217
  20. 2020A Survey on Causal Inferencecited by 202
  21. 2023Inspecting and Editing Knowledge Representations in Language Modelscited by 140
  22. 2023AI Transparency in the Age of LLMs: A Human-Centered Research Roadmapcited by 119
  23. 2021An Empirical Survey of the Effectiveness of Debiasing Techniques for Pre-trained Language Modelscited by 96
  24. 2023Is ChatGPT a Good Causal Reasoner? A Comprehensive Evaluationcited by 69
  25. 2023Improving Activation Steering in Language Models with Mean-Centringcited by 64
  26. 2023Enhancing Recommender Systems with Large Language Model Reasoning Graphscited by 52
  27. 2023Large Language Models are Temporal and Causal Reasoners for Video Question Answeringcited by 51
  28. 2021Graph Infomax Adversarial Learning for Treatment Effect Estimation with Networked Observational Datacited by 42
  29. 2023Leveraging Large Language Models for Pre-trained Recommender Systemscited by 42
  30. 2022Can Prompt Probe Pretrained Language Models? Understanding the Invisible Risks from a Causal Viewcited by 31
  31. 2020Matching in Selective and Balanced Representation Space for Treatment Effects Estimationcited by 30
  32. 2021Defending Pre-trained Language Models from Adversarial Word Substitutions Without Performance Sacrificecited by 30
  33. 2023Intelligent Virtual Assistants with LLM-based Process Automationcited by 26
  34. 2023An Empirical Study of Metrics to Measure Representational Harms in Pre-Trained Language Modelscited by 24
  35. 2023Prompt-augmented Temporal Point Process for Streaming Event Sequencecited by 24
  36. 2024LLM-Guided Multi-View Hypergraph Learning for Human-Centric Explainable Recommendationcited by 19
  37. 2022ROCK: Causal Inference Principles for Reasoning about Commonsense Causalitycited by 16
  38. 2021Word Embeddings via Causal Inference: Gender Bias Reducing and Semantic Information Preservingcited by 11

50 most cited of 59 All 59 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 5; 2025: 6; 2026: 7.

18 papers.

Sort:

By area:

  1. 2024Large Language Model Unlearning via Embedding-Corrupted Promptscited by 112
  2. 2025A Unified Understanding and Evaluation of Steering Methodscited by 40
  3. 2024A Survey on Medical Large Language Models: Technology, Application, Trustworthiness, and Future Directionscited by 34
  4. 2025Correctness Assessment of Code Generated by Large Language Models Using Internal Representationscited by 17
  5. 2025SoK: Evaluating Jailbreak Guardrails for Large Language Modelscited by 17
  6. 2024Attack Atlas: A Practitioner's Perspective on Challenges and Pitfalls in Red Teaming GenAIcited by 14
  7. 2024Swiss Cheese Model for AI Safety: A Taxonomy and Reference Architecture for Multi-Layered Guardrails of Foundation Model Based Agentscited by 13
  8. 2025Standard Benchmarks Fail -- Auditing LLM Agents in Finance Must Prioritize Riskcited by 13
  9. 2026Steering Vector Fields for Context-Aware Inference-Time Control in Large Language Modelscited by 10
  10. 2025Toward a Flexible Framework for Linear Representation Hypothesis Using Maximum Likelihood Estimationcited by 3
  11. 2025Causality Is Key to Understand and Balance Multiple Goals in Trustworthy ML and Foundation Modelscited by 2
  12. 2024Prompt-Consistency Image Generation (PCIG): A Unified Framework Integrating LLMs, Knowledge Graphs, and Controllable Diffusion Modelscited by 1
6 more
  1. 2026ARGORA: Orchestrated Argumentation for Causally Grounded LLM Reasoning and Decision Makingcited by 1
  2. 2026Explainable Token-level Noise Filtering for LLM Fine-tuning Datasetscited by 1
  3. 2026ConsisGuard: Aligning Safety Deliberation with Policy Enforcement in LLM Guardrailscited by 1
  4. 2026A Lightweight Explainable Guardrail for Prompt Safetycited by 0
  5. 2026Auditing Framing-Sensitive Behavioral Instability in Large Language Models for Mental Health Interactionscited by 0
  6. 2026The Safeguard Worked. Is the LLM System Safer?cited by 0