mlsanityData to 13 Oct 2026

AURA: Affordance-Understanding and Risk-aware Alignment Technique for Large Language Models

arXiv 2508.06124 ↗First posted 8 Aug 2025cs.CLcited by 2

On the map: LLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks

Abstract (arXiv)

Present day LLMs face the challenge of managing affordance-based safety risks-situations where outputs inadvertently facilitate harmful actions due to overlooked logical implications. Traditional safety solutions, such as scalar outcome-based reward models, parameter tuning, or heuristic decoding strategies, lack the granularity and proactive nature needed to reliably detect and intervene during subtle yet crucial reasoning steps. Addressing this fundamental gap, we introduce AURA, an innovative, multi-layered framework centered around Process Reward Models (PRMs), providing comprehensive, step level evaluations across logical coherence and safety-awareness. Our framework seamlessly combines introspective self-critique, fine-grained PRM assessments, and adaptive safety-aware decoding to dynamically and proactively guide models toward safer reasoning trajectories. Empirical evidence clearly demonstrates that this approach significantly surpasses existing methods, significantly improving the logical integrity and affordance-sensitive safety of model outputs. This research represents a pivotal step toward safer, more responsible, and contextually aware AI, setting a new benchmark for alignment-sensitive applications.

Key facts. Cited by 2 papers in the map (data through 13 Oct 2026). Placed in Safety alignment and jailbreak attacks, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 25 of its 33 references that are in the map. First posted on arXiv 8 Aug 2025.

Builds on

25 of 33 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025Soteria: Language-Specific Functional Parameter Steering for Multilingual Safety Alignment16 Feb
  2. 2024Controllable Safety Alignment: Inference-Time Adaptation to Diverse Safety Requirements11 Oct
  3. 2023Safe RLHF: Safe Reinforcement Learning from Human Feedback19 Oct

Other references · 22

  1. 2021Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processingcited by 2,229
  2. 2021Calibrate Before Use: Improving Few-Shot Performance of Language Modelscited by 1,376
  3. 2023CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquingcited by 764
  4. 2025The Lessons of Developing Process Reward Models in Mathematical Reasoningcited by 407
  5. 2024ProcessBench: Identifying Process Errors in Mathematical Reasoningcited by 280
  6. 2025Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safetycited by 236
  7. 2021How Many Data Points is a Prompt Worth?cited by 197
  8. 2024Mutual Reasoning Makes Smaller LLMs Stronger Problem-Solverscited by 138
  9. 2025PRMBench: A Fine-grained and Challenging Benchmark for Process-Level Reward Modelscited by 98
  10. 2025Process Reward Models for LLM Agents: Practical Framework and Directionscited by 82
  11. 2024Evaluating Mathematical Reasoning Beyond Accuracycited by 69
  12. 2025RealSafe-R1: Safety-Aligned DeepSeek-R1 without Compromising Reasoning Capabilitycited by 59
10 more
  1. 2025R-PRM: Reasoning-Driven Process Reward Modelingcited by 38
  2. 2024How (un)ethical are instruction-centric responses of LLMs? Unveiling the vulnerabilities of safety guardrails to harmful queriescited by 27
  3. 2025AURORA:Automated Training Framework of Universal Process Reward Models via Ensemble Prompting and Reverse Verificationcited by 22
  4. 2025Subtle Risks, Critical Failures: A Framework for Diagnosing Physical Safety of LLMs for Embodied Decision Makingcited by 20
  5. 2025Athena: Enhancing Multimodal Reasoning with Data-efficient Process Reward Modelscited by 16
  6. 2024Structure Guided Prompt: Instructing Large Language Model in Multi-Step Reasoning by Exploring Graph Structure of the Textcited by 14
  7. 2025From Outcomes to Processes: Guiding PRM Learning from ORM for Inference-Time Alignmentcited by 9
  8. 2025Error Typing for Smarter Rewards: Improving Process Reward Models with Error-Aware Hierarchical Supervisioncited by 8
  9. 2025SafeTy Reasoning Elicitation Alignment for Multi-Turn Dialoguescited by 7
  10. 2025Scalable Oversight for Superhuman AI via Recursive Self-Critiquingcited by 4

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Towards Robust Process Reward Modeling via Noise-aware Learningcited by 2
  2. 2026MortalMATH: Evaluating the Conflict Between Reasoning Objectives and Emergency Contextscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Vulnerability Mitigation for Safety-Aligned Language Models via Debiasing
  2. 2024Towards Understanding Safety Alignment: A Mechanistic Perspective from Safety Neurons
  3. 2026The Causal Impact of Tool Affordance on Safety Alignment in LLM Agents
  4. 2025Picky LLMs and Unreliable RMs: An Empirical Study on Safety Alignment after Instruction Tuning
  5. 2025SAFER: Advancing Safety Alignment via Efficient Ex-Ante Reasoning
  6. 2025Beyond Safe Answers: A Benchmark for Evaluating True Risk Awareness in Large Reasoning Models
  7. 2025Read the Scene, Not the Script: Outcome-Aware Safety for LLMs
  8. 2026SafeTune: A Unified Faithful Library for Auditing and Repairing Safety Drift in Fine-Tuned LLMs