arXiv:2609.36254v1 Announce Type: new Abstract: Large Reasoning Models (LRMs) are commonly trained with reinforcement learning (RL) to improve their generation of chain-of-thought (CoT) reasoning bef...
The paper introduces SaLT-DPO, a Segment‑aware Listwise Target DPO method designed to improve safety in Large Reasoning Models (LRMs). It decomposes responses into reasoning and answer segments, scores each segment’s safety, and aligns length‑normalized rewards with soft target distributions over multiple candidates. The approach also incorporates joint safety coherence regularization and utility anchoring to reduce unsafe content while maintaining compliance and reasoning performance.
The paper "Rethinking GRPO for Large Reasoning Model Safety Alignment" identifies two failure modes in Group Relative Policy Optimization—advantage collapse and unsafe diversification—that weaken safety learning in large reasoning models. It proposes the Beacon framework, which includes Safety Signal Recovery (SSR) and Safety-Utility Decoupled Alignment (SUDA), to restore effective safety signals and mitigate over-refusal. Experiments on Qwen3 and DeepSeek-R1-Distill models across eight benchmarks demonstrate that Beacon improves safety performance by up to 32.83 percentage points while keeping over-refusal low.
arXiv:2608. 09542v1 Announce Type: cross Abstract: Large reasoning models (LRMs) achieve remarkable success on complex tasks but remain vulnerable to harmful prompts that induce unsafe outputs.
ReasoningShield introduces a benchmark and lightweight models for detecting unsafe content in the intermediate reasoning steps of Large Reasoning Models (LRMs). The benchmark contains 9.2K annotated (Query, CoT) pairs across 10 risk categories, 3 safety levels, and 8 LRMs, and the proposed models achieve 91.8% F1, outperforming existing moderation tools by significant margins. The study highlights that current moderation tools degrade by up to 35.2% on Chain-of-Thought reasoning compared to traditional answer moderation.
Large Reasoning Models (LRMs) use explicit chain‑of‑thought reasoning and large context windows to perform complex tasks, but these features create new attack surfaces. The paper introduces SRCF, an attack that steers LRMs by prepending counter‑aligned few‑shot conversations with explicit CoT traces, causing unsafe outputs on harmful queries and unwarranted refusals on benign ones, without needing model internals. To counter this, the authors propose ARCF, a post‑training defense that exposes models to counter‑aligned conversational contexts while enforcing aligned targets, improving safety and helpfulness without harming utility.
arXiv:2607. 06326v1 Announce Type: new Abstract: Large language models deployed in open-world applications require safety guardrails that are both robust to complex risks and efficient enough for low-latency runtime moderation.
arXiv:2506. 07031v5 Announce Type: replace-cross Abstract: Emerging Large Reasoning Models (LRMs) consistently excel in mathematical and reasoning tasks, showcasing remarkable capabilities.
arXiv:2609.23587v1 Announce Type: new Abstract: Large reasoning models (LRMs) incur high inference costs, often mitigated by efficiency techniques like quantization and pruning. However, the impact o...
Large reasoning models generate chain-of-thought (CoT) before answering but struggle with safety alignment and can be misled by flawed premises. The paper introduces RECAP, a reinforcement learning approach that trains models to override flawed reasoning paths and produce safe, helpful responses without extra training cost. RECAP improves safety, jailbreak robustness, and reduces overrefusal while preserving core reasoning abilities and inference token budget.
arXiv:2608.23497v1 Announce Type: new Abstract: Reasoning-Induced Misalignment, where fine-tuning on reasoning data containing no harmful content, including mathematics, code, and problem-solving wit...
The paper investigates the reliability of rule- and model-based verifiers used in reinforcement learning with verifiable reward (RLVR) for mathematical reasoning. It finds that rule-based verifiers often miss equivalent answers in different formats, causing false negatives that degrade RL performance as models improve. Model-based verifiers achieve higher static accuracy but become vulnerable to reward hacking during RL, misclassifying certain response patterns as correct after fine-tuning.