NExT-Guard: Training-Free Streaming Safeguard without Token-Level Labels

• URL: http://arxiv.org/abs/2603.02219v1
• Date: Tue, 10 Feb 2026 08:54:10 GMT
• Title: NExT-Guard: Training-Free Streaming Safeguard without Token-Level Labels
• Abstract summary: NExT-Guard is a training-free framework that achieves streaming safeguards by monitoring interpretable latent features from Sparse Autoencoders (SAEs)<n>It uses pretrained SAEs from publicly available base LLMs, enabling flexible, low-cost deployment without token-level supervision.<n> Experimental results show that NExT-Guard outperforms both post-hoc and streaming safeguards based on supervised training.
• Score: 66.70287556682105
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models are increasingly deployed in streaming scenarios, rendering conventional post-hoc safeguards ineffective as they fail to interdict unsafe content in real-time. While streaming safeguards based on token-level supervised training could address this, they necessitate expensive annotations and suffer from severe overfitting. In this work, we challenge the paradigm that streaming safety must rely on token-level supervised training. Instead, it is an inherent capability of well-trained post-hoc safeguards, as they already encode token-level risk signals in hidden representations. Hence, we introduce NExT-Guard, a training-free framework that achieves streaming safeguards by monitoring interpretable latent features from Sparse Autoencoders (SAEs). It uses pretrained SAEs from publicly available base LLMs, enabling flexible, low-cost deployment without token-level supervision. Experimental results show that NExT-Guard outperforms both post-hoc and streaming safeguards based on supervised training, with superior robustness across models, SAE variants, and risk scenarios. These results make NExT-Guard a universal and scalable paradigm for real-time safety, accelerating the practical deployment of streaming safeguards.

Related papers

• LatentGuard: Efficient and Inspectable Latent Reasoning for LLM Safeguards [21.146106913383868]
  Reasoning-based guard models improve LLM safeguards, but decoding explicit rationales for every interaction makes them costly to deploy.<n>We propose LatentGuard, an efficient and inspectable safeguard framework that brings continuous latent reasoning to guard models.
  arXiv  Detail & Related papers  (2026-08-04T15:47:41Z)
• BraveGuard: From Open-World Threats to Safer Computer-Use Agents [68.38397514360227]
  BraveGuard is a self-evolving defense framework for training guard models from open-world threat signals and realistic agent trajectories.<n>We instantiate BraveGuard by training multiple guard backbones, including Qwen3-Guard and Llama-Guard variants, and evaluate the resulting guards on trajectory-level agent-safety benchmarks.
  arXiv  Detail & Related papers  (2026-05-31T11:16:18Z)
• Steering Externalities: Benign Activation Steering Unintentionally Increases Jailbreak Risk for Large Language Models [62.16655896700062]
  Activation steering is a technique to enhance the utility of Large Language Models (LLMs)<n>We show that it unintentionally introduces critical and under-explored safety risks.<n>Experiments reveal that these interventions act as a force multiplier, creating new vulnerabilities to jailbreaks and increasing attack success rates to over 80% on standard benchmarks.
  arXiv  Detail & Related papers  (2026-02-03T12:32:35Z)
• One Leak Away: How Pretrained Model Exposure Amplifies Jailbreak Risks in Finetuned LLMs [8.493938164845249]
  Finetuning pretrained large language models (LLMs) has become the standard paradigm for developing downstream applications.<n>We investigate whether finetuned LLMs inherit jailbreak vulnerabilities from their pretrained sources.<n>We propose the Probe-Guided Projection (PGP) attack, which steers optimization toward transferability-relevant directions.
  arXiv  Detail & Related papers  (2025-12-14T07:48:44Z)
• Efficient Switchable Safety Control in LLMs via Magic-Token-Guided Co-Training [1.5349686675266894]
  Current methods for content safety in Large Language Models (LLMs) rely on multi-stage training pipelines.<n>We propose a unified co-training framework that efficiently integrates multiple safety behaviors.<n>We show that our method matches the safety alignment quality of SFT+DPO, with our 8B model notably surpassing DeepSeek-R1 (671B) in safety performance.
  arXiv  Detail & Related papers  (2025-08-12T02:39:33Z)
• Secure Tug-of-War (SecTOW): Iterative Defense-Attack Training with Reinforcement Learning for Multimodal Model Security [63.41350337821108]
  We propose Secure Tug-of-War (SecTOW) to enhance the security of multimodal large language models (MLLMs)<n>SecTOW consists of two modules: a defender and an auxiliary attacker, both trained iteratively using reinforcement learning (GRPO)<n>We show that SecTOW significantly improves security while preserving general performance.
  arXiv  Detail & Related papers  (2025-07-29T17:39:48Z)
• SafeSwitch: Steering Unsafe LLM Behavior via Internal Activation Signals [51.49737867797442]
  Large language models (LLMs) exhibit exceptional capabilities across various tasks but also pose risks by generating harmful content.<n>We show that LLMs can similarly perform internal assessments about safety in their internal states.<n>We propose SafeSwitch, a framework that regulates unsafe outputs by utilizing the prober-based internal state monitor.
  arXiv  Detail & Related papers  (2025-02-03T04:23:33Z)
• CoreGuard: Safeguarding Foundational Capabilities of LLMs Against Model Stealing in Edge Deployment [66.72332011814183]
  CoreGuard is a computation- and communication-efficient protection method for proprietary large language models (LLMs) deployed on edge devices.<n> CoreGuard employs an efficient protection protocol to reduce computational overhead and minimize communication overhead via a propagation protocol.
  arXiv  Detail & Related papers  (2024-10-16T08:14:24Z)
• LLMs know their vulnerabilities: Uncover Safety Gaps through Natural Distribution Shifts [88.96201324719205]
  Safety concerns in large language models (LLMs) have gained significant attention due to their exposure to potentially harmful data during pre-training.<n>We identify a new safety vulnerability in LLMs, where seemingly benign prompts, semantically related to harmful content, can bypass safety mechanisms.<n>We introduce a novel attack method, textitActorBreaker, which identifies actors related to toxic prompts within pre-training distribution.
  arXiv  Detail & Related papers  (2024-10-14T16:41:49Z)
• Evaluating Defences against Unsafe Feedback in RLHF [26.872318173182414]
  This paper looks at learning from unsafe feedback with reinforcement learning.<n>We find that safety-aligned LLMs easily explore unsafe action spaces via generating harmful text.<n>In order to protect against this vulnerability, we adapt a number of both "implict" and "explicit" harmful fine-tuning defences.
  arXiv  Detail & Related papers  (2024-09-19T17:10:34Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.