Safin-1: Safety from Within through Memory-Native State Evolution

• URL: http://arxiv.org/abs/2609.00092v1
• Date: Mon, 31 Aug 2026 13:48:48 GMT
• Title: Safin-1: Safety from Within through Memory-Native State Evolution
• Abstract summary: Safin-1 is a family of foundation models realizing this principle through memory routing and state evolution.<n> Safin-1 is built on Memory-Anchor Routing across Context History (MARCH), a network architecture that maintains structured memory states.<n>We investigate this interface on downstream safety tasks through a Safety State, demonstrating effective state-based adaptation with substantial safety improvements.
• Score: 51.221955257147414
• License: http://creativecommons.org/publicdomain/zero/1.0/
• Abstract: Long-horizon complex tasks require foundation models to accumulate information, maintain internal states, and adapt over extended interactions. Safety should be an intrinsic property of the model itself, rather than a behavioral constraint relying solely on external safeguards or post-hoc alignment such as supervised fine-tuning. This motivates Safety from Within, where safety-relevant capabilities are represented and invoked through the model's native computation. We present Safin-1, a family of foundation models realizing this principle through memory routing and state evolution. Safin-1 is built on Memory-Anchor Routing across Context History (MARCH), a network architecture that maintains structured memory states and selectively retrieves relevant historical information through content-conditioned routing. It supports test-time adaptation of persistent capability states without repeatedly modifying the backbone, enabling controlled specialization over a shared foundation. We investigate this interface on downstream safety tasks through a Safety State, demonstrating effective state-based adaptation with substantial safety improvements. More broadly, the routed-state interface unifies contextual memory and persistent capability adaptation within the model's native computation, reframing memory from a passive record of prior context into an active substrate for maintaining and evolving model behavior. Evaluations across general capabilities, long-context understanding, retrieval, and efficiency further validate Safin-1. These findings provide a path toward safety as a state-native and adaptively maintainable capability. This work is only an initial architectural exploration of Safety from Within, and substantial further work is needed to realize this broader vision.

Related papers

• Do Models Share Safety Representations? Cross-Model Steering for Safe Visual Generation [52.122731171289665]
  We introduce the first framework for cross-model safety steering.<n>Our pipeline never accesses unsafe data on the target side.<n>We evaluate our approach in text-to-image and text-to-video generation across diverse source-target model pairs.
  arXiv  Detail & Related papers  (2026-06-03T18:00:04Z)
• RefusalGuard: Geometry-Preserving Fine-Tuning for Safety in LLMs [5.100622189286672]
  Fine-tuning safety-aligned language models for downstream tasks can lead to substantial degradation of refusal behavior.<n>We introduce REFUSALGUARD, a representation-level fine-tuning framework that preserves safety-relevant structure during model adaptation.
  arXiv  Detail & Related papers  (2026-05-03T14:48:18Z)
• PoSafeNet: Safe Learning with Poset-Structured Neural Nets [49.854863600271614]
  existing approaches often enforce multiple safety constraints uniformly or via fixed priority orders, leading to infeasibility and brittle behavior.<n>We formalize this setting as poset-structured safety, modeling safety constraints as a partially ordered set and treating safety composition as a structural property of the policy class.<n>Building on this formulation, we propose PoSafeNet, a differentiable neural safety layer that enforces safety via sequential closed-form projection.
  arXiv  Detail & Related papers  (2026-01-29T22:03:32Z)
• SafeRedir: Prompt Embedding Redirection for Robust Unlearning in Image Generation Models [67.84174763413178]
  We introduce SafeRedir, a lightweight inference-time framework for robust unlearning via prompt embedding redirection.<n>We show that SafeRedir achieves effective unlearning capability, high semantic and perceptual preservation, robust image quality, and enhanced resistance to adversarial attacks.
  arXiv  Detail & Related papers  (2026-01-13T15:01:38Z)
• SafeCast: Risk-Responsive Motion Forecasting for Autonomous Vehicles [12.607007386467329]
  We present SafeCast, a risk-responsive motion forecasting model.<n>It integrates safety-aware decision-making with uncertainty-aware adaptability.<n>Our model achieves state-of-the-art (SOTA) accuracy while maintaining a lightweight architecture and low inference latency.
  arXiv  Detail & Related papers  (2025-03-28T15:38:21Z)
• Learning Vision-Based Neural Network Controllers with Semi-Probabilistic Safety Guarantees [24.650302053973142]
  We introduce a novel semi-probabilistic verification framework that integrates reachability analysis with conditional generative adversarial networks.<n>Next, we develop a gradient-based training approach that employs a novel safety loss function, safety-aware data-sampling strategy, and curriculum learning.<n> Empirical evaluations in X-Plane 11 airplane landing simulation, CARLA-simulated autonomous lane following, and F1Tenth lane following in a visually-rich miniature environment demonstrate the effectiveness of our method in achieving formal safety guarantees.
  arXiv  Detail & Related papers  (2025-02-28T21:16:42Z)
• Safe Reinforcement Learning with Learned Non-Markovian Safety Constraints [15.904640266226023]
  We design a safety model that performs credit assignment to assess contributions of partial state-action trajectories on safety. We derive an effective algorithm for optimizing a safe policy using the learned safety model. We devise a method to dynamically adapt the tradeoff coefficient between safety reward and safety compliance.
  arXiv  Detail & Related papers  (2024-05-05T17:27:22Z)
• Evaluating Model-free Reinforcement Learning toward Safety-critical Tasks [70.76757529955577]
  This paper revisits prior work in this scope from the perspective of state-wise safe RL. We propose Unrolling Safety Layer (USL), a joint method that combines safety optimization and safety projection. To facilitate further research in this area, we reproduce related algorithms in a unified pipeline and incorporate them into SafeRL-Kit.
  arXiv  Detail & Related papers  (2022-12-12T06:30:17Z)
• Recursively Feasible Probabilistic Safe Online Learning with Control Barrier Functions [60.26921219698514]
  We introduce a model-uncertainty-aware reformulation of CBF-based safety-critical controllers. We then present the pointwise feasibility conditions of the resulting safety controller. We use these conditions to devise an event-triggered online data collection strategy.
  arXiv  Detail & Related papers  (2022-08-23T05:02:09Z)

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.