# Robust Critics: Defending LLMs Against Multi-Turn Attacks

arXiv: https://arxiv.org/abs/2607.20472 · HTML: https://mlsanity.com/paper/2607.20472/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 52 of its 62 references that are in the map.
- First posted on arXiv 24 May 2026.

## Placed in

- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 30 of 52)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor](https://mlsanity.com/paper/1801.01290/) (2018)
- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [Constitutional AI: Harmlessness from AI Feedback](https://mlsanity.com/paper/2212.08073/) (2022)
- [High-Dimensional Continuous Control Using Generalized Advantage Estimation](https://mlsanity.com/paper/1506.02438/) (2015)
- [HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal](https://mlsanity.com/paper/2402.04249/) (2024)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [Tulu 3: Pushing Frontiers in Open Language Model Post-Training](https://mlsanity.com/paper/2411.15124/) (2024)
- [Safe RLHF: Safe Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2310.12773/) (2023)
- [XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Models](https://mlsanity.com/paper/2308.01263/) (2023)
- [SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks](https://mlsanity.com/paper/2310.03684/) (2023)
- [Robust Adversarial Reinforcement Learning](https://mlsanity.com/paper/1703.02702/) (2017)
- [Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack](https://mlsanity.com/paper/2404.01833/) (2024)
- [Hierarchical Deep Reinforcement Learning: Integrating Temporal Abstraction and Intrinsic Motivation](https://mlsanity.com/paper/1604.06057/) (2016)
- [OR-Bench: An Over-Refusal Benchmark for Large Language Models](https://mlsanity.com/paper/2405.20947/) (2024)
- [Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications](https://mlsanity.com/paper/2402.05162/) (2024)
- [ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RL](https://mlsanity.com/paper/2402.19446/) (2024)
- [Evaluating Large Language Models in Theory of Mind Tasks](https://mlsanity.com/paper/2302.02083/) (2023)
- [WildBench: Benchmarking LLMs with Challenging Tasks from Real Users in the Wild](https://mlsanity.com/paper/2406.04770/) (2024)
- [Adversarial Policies: Attacking Deep Reinforcement Learning](https://mlsanity.com/paper/1905.10615/) (2019)
- [Controlled Decoding from Language Models](https://mlsanity.com/paper/2310.17022/) (2023)
- [X-Teaming: Multi-Turn Jailbreaks and Defenses with Adaptive Multi-Agents](https://mlsanity.com/paper/2504.13203/) (2025)
- [On Prompt-Driven Safeguarding for Large Language Models](https://mlsanity.com/paper/2401.18018/) (2024)
- [Efficient Adversarial Training in LLMs with Continuous Attacks](https://mlsanity.com/paper/2405.15589/) (2024)
- [The Art of Saying No: Contextual Noncompliance in Language Models](https://mlsanity.com/paper/2407.12043/) (2024)
- [Reward-Augmented Decoding: Efficient Controlled Text Generation With a Unidirectional Reward Model](https://mlsanity.com/paper/2310.09520/) (2023)
- [FalseReject: A Resource for Improving Contextual Safety and Mitigating Over-Refusals in LLMs via Structured Reasoning](https://mlsanity.com/paper/2505.08054/) (2025)

## Abstract (arXiv)

When a user asks a language model something harmful, is it a genuine attack or a misunderstood but well-meaning question? This ambiguity is one of the central challenges of LLM safety. A model that assumes the worst harms legitimate users; one that assumes the best is easily exploited. The problem is compounded in multi-turn dialogue, where an attacker's true intent may only reveal itself gradually across many exchanges, yet existing safety frameworks apply a contextual bandit treatment, ignoring the trajectory of the conversation. To that end, we propose Dialogue Critic Guided Sampling (DCGS), a framework that addresses this by inferring user intent at every turn of dialogue. Instead of applying a fixed rule about what is or is not safe, DCGS learns what the user's intent is likely to be based on the full conversational history and generates responses accordingly. Formally, we model adversarial dialogue as a Markov Decision Process and learn value and regret-based critics at both the individual token and utterance (full response) levels, scoring candidate responses via an action-value critic. We prove that this inference-time reweighting approximates exponential tilting of the base policy, guaranteeing improvement in expected return for any finite candidate pool, a property that group-relative objectives do not exhibit. Evaluated on CARES-18k, WildJailbreak, Redbench, and Harmbench, DCGS outperforms strong robust baselines and frontier models on adversarial dialogue tasks. DCGS also transfers to frontier models, improving their robustness without fine-tuning.