Safety Alignment

Benchmarks archive 2025-07-28

No benchmark for this task in the archive.

Libraries

Not in the archive: the export carries no per-task library table, so there is nothing to show at snapshot 2025-07-28.

Datasets archive 2025-07-28

1 dataset whose archive record lists this task, ordered by the archive's paper count.

Subtasks archive 2025-07-28

No subtask under this task in the archive's task tree.

Most implemented papers archive 2025-07-28

30 shown of 134 papers with code (288 tagged with this task in all), ordered by repositories listed in the archive, not by stars (the archive holds no stars, so PwC's “Social” and “Latest” sorts cannot be reproduced). Papers without a page here are shown as plain text.

• The Hidden Dimensions of LLM Alignment: A Multi-Dimensional Safety Analysis
  13 Feb 2025 3 repositories listed
  We then measure how different directions promote or suppress the dominant direction, showing the important role of secondary directions in shaping the model's refusal representation.
• Antidote: Post-fine-tuning Safety Alignment for Large Language Models against Harmful Fine-tuning
  18 Aug 2024 3 repositories listed Syntology ran 3 of 5 samples · 2 unverified
  To this end, we propose Antidote, a post-fine-tuning stage solution, which remains \textbf{\textit{agnostic to the training hyper-parameters in the fine-tuning stage}}.
• Derail Yourself: Multi-turn LLM Jailbreak Attack through Self-discovered Clues
  14 Oct 2024 2 repositories listed Syntology ran 2 of 3 samples · 1 unverified · 3 pointer-only (licence)
  This study exposes the safety vulnerabilities of Large Language Models (LLMs) in multi-turn interactions, where malicious users can obscure harmful intents across several queries.
• Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey
  26 Sep 2024 2 repositories listed Syntology ran 6 of 9 samples · 3 unverified
  To clear up concern, this paper provide a comprehensive overview to three aspects of harmful fine-tuning: attacks setting, defense design and evaluation methodology.
• CodeAttack: Revealing Safety Generalization Challenges of Large Language Models via Code Completion
  12 Mar 2024 2 repositories listed Syntology ran 2 of 2 samples · 0 unverified
  The rapid advancement of Large Language Models (LLMs) has brought about remarkable generative capabilities but also raised concerns about their potential misuse.
• FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Prompts
  9 Nov 2023 2 repositories listed Syntology ran 1 of 1 samples · 0 unverified · 1 pointer-only (licence)
  Large Vision-Language Models (LVLMs) signify a groundbreaking paradigm shift within the Artificial Intelligence (AI) community, extending beyond the capabilities of Large Language Models (LLMs) by assimilating…
• Beyond One-Preference-Fits-All Alignment: Multi-Objective Direct Preference Optimization
  5 Oct 2023 2 repositories listed
  A single language model, even when aligned with labelers through reinforcement learning from human feedback (RLHF), may not suit all human preferences.
• Red-Teaming Large Language Models using Chain of Utterances for Safety-Alignment
  18 Aug 2023 2 repositories listed
  In this work, we propose a new safety evaluation benchmark RED-EVAL that carries out red-teaming.
• The Devil behind the mask: An emergent safety vulnerability of Diffusion LLMs
  15 Jul 2025 1 repository listed Syntology ran 11 of 17 samples · 6 unverified
  Diffusion-based large language models (dLLMs) have recently emerged as a powerful alternative to autoregressive LLMs, offering faster inference and greater interactivity via parallel decoding and bidirectional modeling.
• Safe Pruning LoRA: Robust Distance-Guided Pruning for Safety Alignment in Adaptation of LLMs
  21 Jun 2025 1 repository listed Syntology ran 0 of 8 samples · 8 unverified · 8 pointer-only (licence)
  Fine-tuning Large Language Models (LLMs) with Low-Rank Adaptation (LoRA) enhances adaptability while reducing computational costs.
• Probe before You Talk: Towards Black-box Defense against Backdoor Unalignment for Large Language Models
  19 Jun 2025 1 repository listed
  Backdoor unalignment attacks against Large Language Models (LLMs) enable the stealthy compromise of safety alignment using a hidden trigger while evading normal safety auditing.
• Probing the Robustness of Large Language Models Safety to Latent Perturbations
  19 Jun 2025 1 repository listed
  Consequently, small shifts in hidden activations can re-trigger harmful behaviors embedded in the latent space.
• Mitigating Safety Fallback in Editing-based Backdoor Injection on LLMs
  16 Jun 2025 1 repository listed
  Large language models (LLMs) have shown strong performance across natural language tasks, but remain vulnerable to backdoor attacks.
• Monitoring Decomposition Attacks in LLMs with Lightweight Sequential Monitors
  12 Jun 2025 1 repository listed
  We show that a carefully prompt engineered lightweight monitor achieves a 93% defense success rate, beating reasoning models like o3 mini as a monitor.
• DAVSP: Safety Alignment for Large Vision-Language Models via Deep Aligned Visual Safety Prompt
  11 Jun 2025 1 repository listed
  Extensive experiments across five benchmarks on two representative LVLMs demonstrate that DAVSP effectively resists malicious queries while preserving benign input utility.
• RSafe: Incentivizing proactive reasoning to build robust and adaptive LLM safeguards
  9 Jun 2025 1 repository listed
  Large Language Models (LLMs) continue to exhibit vulnerabilities despite deliberate safety alignment efforts, posing significant risks to users and society.
• Chasing Moving Targets with Online Self-Play Reinforcement Learning for Safer Language Models
  9 Jun 2025 1 repository listed Syntology ran 6 of 17 samples · 11 unverified · 1 pointer-only (licence)
  Conventional language model (LM) safety alignment relies on a reactive, disjoint procedure: attackers exploit a static model, followed by defensive fine-tuning to patch exposed vulnerabilities.
• BitBypass: A New Direction in Jailbreaking Aligned Large Language Models with Bitstream Camouflage
  3 Jun 2025 1 repository listed
  The inherent risk of generating harmful and unsafe content by Large Language Models (LLMs), has highlighted the need for their safety alignment.
• DiaBlo: Diagonal Blocks Are Sufficient For Finetuning
  3 Jun 2025 1 repository listed Syntology ran 2 of 3 samples · 1 unverified · 3 pointer-only (licence)
  Finetuning is a critical step for adapting large language models (LLMs) to domain-specific downstream tasks.
• TRIDENT: Enhancing Large Language Model Safety with Tri-Dimensional Diversified Red-Teaming Data Synthesis
  30 May 2025 1 repository listed
  Large Language Models (LLMs) excel in various natural language processing tasks but remain vulnerable to generating harmful content or being exploited for malicious purposes.
• AgentAlign: Navigating Safety Alignment in the Shift from Informative to Agentic Large Language Models
  29 May 2025 1 repository listed
  The acquisition of agentic capabilities has transformed LLMs from "knowledge providers" to "action executors", a trend that while expanding LLMs' capability boundaries, significantly increases their susceptibility to…
• OVERT: A Benchmark for Over-Refusal Evaluation on Text-to-Image Models
  27 May 2025 1 repository listed Syntology ran 0 of 1 samples · 1 unverified
  Text-to-Image (T2I) models have achieved remarkable success in generating visual content from text inputs.
• Lifelong Safety Alignment for Language Models
  26 May 2025 1 repository listed
  LLMs have made impressive progress, but their growing capabilities also expose them to highly flexible jailbreaking attacks designed to bypass safety alignment.
• Beyond Safe Answers: A Benchmark for Evaluating True Risk Awareness in Large Reasoning Models
  26 May 2025 1 repository listed
  Despite the remarkable proficiency of \textit{Large Reasoning Models} (LRMs) in handling complex reasoning tasks, their reliability in safety-critical scenarios remains uncertain.
• VSCBench: Bridging the Gap in Vision-Language Model Safety Calibration
  26 May 2025 1 repository listed
  In this paper, we introduce the concept of safety calibration, which systematically addresses both undersafety and oversafety.
• One Model Transfer to All: On Robust Jailbreak Prompts Generation against LLMs
  23 May 2025 1 repository listed
  Investigating these attacks is crucial for uncovering model vulnerabilities.
• Mitigating Fine-tuning Risks in LLMs via Safety-Aware Probing Optimization
  22 May 2025 1 repository listed
  The significant progress of large language models (LLMs) has led to remarkable achievements across numerous applications.
• DuFFin: A Dual-Level Fingerprinting Framework for LLMs IP Protection
  22 May 2025 1 repository listed
  Large language models (LLMs) are considered valuable Intellectual Properties (IP) for legitimate owners due to the enormous computational cost of training.
• MPO: Multilingual Safety Alignment via Reward Gap Optimization
  22 May 2025 1 repository listed Syntology ran 2 of 3 samples · 1 unverified
  Large language models (LLMs) have become increasingly central to AI applications worldwide, necessitating robust multilingual safety alignment to ensure secure deployment across diverse linguistic contexts.
• MTSA: Multi-turn Safety Alignment for LLMs through Multi-round Red-teaming
  22 May 2025 1 repository listed
  In the adversarial iterative optimization stage, the red-team model and the target model continuously improve their respective capabilities in interaction.

Syntology lines on 11 of the papers shown; no Syntology record for the others (a paper without an arXiv id cannot be joined to the graph, and absence from the graph layer is not a recorded non-run). “Ran” means the sample executed on a synthesized fixture, not that the paper's result was reproduced. Read from the graph 2026-09-24.