As large language models (LLMs) are deployed in safety-critical settings, it is essential to ensure that their responses comply with safety standards. Prior research has revealed that LLMs often fail to grasp the notion of safe behaviors, resulting in either unjustified refusals to harmless prompts or the generation of harmful content. While substantial efforts have been made to improve their robustness, existing defenses often rely on costly fine-tuning of model parameters or employ suboptimal heuristic techniques. In this work, we take a novel approach to safeguard LLMs by learning to adapt the system prompts in instruction-tuned LLMs. While LLMs are typically pre-trained to follow a fixed system prompt, we investigate the impact of tailoring the system prompt to each specific user input on the safety of the responses. To this end, we propose $\textbf{Sysformer}$, a trans$\textbf{former}$ model that updates an initial $\textbf{sys}$tem prompt to a more robust system prompt in the LLM input embedding space while attending to the user prompt. While keeping the LLM parameters frozen, the Sysformer is trained to refuse to respond to a set of harmful prompts while responding ideally to a set of safe ones. Through extensive experiments on $5$ LLMs from different families and $2$ recent benchmarks, we demonstrate that Sysformer can significantly enhance the robustness of LLMs, leading to upto $80\%$ gain in the refusal rate on harmful prompts while enhancing the compliance with the safe prompts by upto $90\%$. Results also generalize well to sophisticated jailbreaking attacks, making LLMs upto $100\%$ more robust against different attack strategies. We hope our findings lead to cheaper safeguarding of LLMs and motivate future investigations into designing variable system prompts.

As Vision-Language Models (VLMs) demonstrate increasing capabilities across real-world applications such as code generation and chatbot assistance, ensuring their safety has become paramount. Unlike traditional Large Language Models (LLMs), VLMs face unique vulnerabilities due to their multimodal nature, allowing adversaries to modify visual or textual inputs to bypass safety guardrails and trigger the generation of harmful content. Through systematic analysis of VLM behavior under attack, we identify a novel phenomenon termed ``delayed safety awareness''. Specifically, we observe that safety-aligned VLMs may initially be compromised to produce harmful content, but eventually recognize the associated risks and attempt to self-correct. This pattern suggests that VLMs retain their underlying safety awareness but experience a temporal delay in their activation. Building on this insight, we hypothesize that VLMs' safety awareness can be proactively reactivated through carefully designed prompts. To this end, we introduce ``The Safety Reminder'', a soft prompt tuning approach that optimizes learnable prompt tokens, which are periodically injected during the text generation process to enhance safety awareness, effectively preventing harmful content generation. Additionally, our safety reminder only activates when harmful content is detected, leaving normal conversations unaffected and preserving the model's performance on benign tasks. Through comprehensive evaluation across three established safety benchmarks and one adversarial attacks, we demonstrate that our approach significantly reduces attack success rates while maintaining model utility, offering a practical solution for deploying safer VLMs in real-world applications.

The integration of large language models (LLMs) into electronic design automation (EDA) has significantly advanced the field, offering transformative benefits, particularly in register transfer level (RTL) code generation and understanding. While previous studies have demonstrated the efficacy of fine-tuning LLMs for these generation-based tasks, embedding-based tasks, which are equally critical to EDA workflows, have been largely overlooked. These tasks, including natural language code search, RTL code functionality equivalence checking, and performance prediction, are essential for accelerating and optimizing the hardware design process. To address this gap, we present DeepRTL2, a family of versatile LLMs that unifies both generation- and embedding-based tasks related to RTL. By simultaneously tackling a broad range of tasks, DeepRTL2 represents the first model to provide a comprehensive solution to the diverse challenges in EDA. Through extensive experiments, we show that DeepRTL2 achieves state-of-the-art performance across all evaluated tasks.

Artificial intelligence (AI) has evolved into an ecosystem of specialized "species," each with unique strengths. We analyze two: DeepSeek-V3, a 671-billion-parameter Mixture of Experts large language model (LLM) exemplifying scale-driven generality, and NetraAI, a dynamical system-based framework engineered for stability and interpretability on small clinical trial datasets. We formalize NetraAI's foundations, combining contraction mappings, information geometry, and evolutionary algorithms to identify predictive patient cohorts. Features are embedded in a metric space and iteratively contracted toward stable attractors that define latent subgroups. A pseudo-temporal embedding and long-range memory enable exploration of higher-order feature interactions, while an internal evolutionary loop selects compact, explainable 2-4-variable bundles ("Personas"). To guide discovery, we introduce an LLM Strategist as a meta-evolutionary layer that observes Persona outputs, prioritizes promising variables, injects domain knowledge, and assesses robustness. This two-tier architecture mirrors the human scientific process: NetraAI as experimentalist, the LLM as theorist, forming a self-improving loop. In case studies (schizophrenia, depression, pancreatic cancer), NetraAI uncovered small, high-effect-size subpopulations that transformed weak baseline models (AUC ~0.50-0.68) into near-perfect classifiers using only a few features. We position NetraAI at the intersection of dynamical systems, information geometry, and evolutionary learning, aligned with emerging concept-level reasoning paradigms such as LeCun's Joint Embedding Predictive Architecture (JEPA). By prioritizing reliable, explainable knowledge, NetraAI offers a new generation of adaptive, self-reflective AI to accelerate clinical discovery.

Data plays the most prominent role in how language models acquire skills and knowledge. The lack of massive, well-organized pre-training datasets results in costly and inaccessible data pipelines. We present Essential-Web v1.0, a 24-trillion-token dataset in which every document is annotated with a twelve-category taxonomy covering topic, format, content complexity, and quality. Taxonomy labels are produced by EAI-Distill-0.5b, a fine-tuned 0.5b-parameter model that achieves an annotator agreement within 3% of Qwen2.5-32B-Instruct. With nothing more than SQL-style filters, we obtain competitive web-curated datasets in math (-8.0% relative to SOTA), web code (+14.3%), STEM (+24.5%) and medical (+8.6%). Essential-Web v1.0 is available on HuggingFace: https://huggingface.co/datasets/EssentialAI/essential-web-v1.0

We study the process through which reasoning models trained with reinforcement learning on verifiable rewards (RLVR) can learn to solve new problems. We find that RLVR drives performance in two main ways: (1) by compressing pass@$k$ into pass@1 and (2) via "capability gain" in which models learn to solve new problems that they previously could not solve even at high $k$. We find that while capability gain exists across model scales, learning to solve new problems is primarily driven through self-distillation. We demonstrate these findings across model scales ranging from 0.5B to 72B parameters on >500,000 reasoning problems with prompts and verifiable final answers across math, science, and code domains. We further show that we can significantly improve pass@$k$ rates by leveraging natural language guidance for the model to consider within context while still requiring the model to derive a solution chain from scratch. Based of these insights, we derive $\text{Guide}$ -- a new class of online training algorithms. $\text{Guide}$ adaptively incorporates hints into the model's context on problems for which all rollouts were initially incorrect and adjusts the importance sampling ratio for the "off-policy" trajectories in order to optimize the policy for contexts in which the hints are no longer present. We describe variants of $\text{Guide}$ for GRPO and PPO and empirically show that Guide-GRPO on 7B and 32B parameter models improves generalization over its vanilla counterpart with up to 4$\%$ macro-average improvement across math benchmarks. We include careful ablations to analyze $\text{Guide}$'s components and theoretically analyze Guide's learning efficiency.

The IPARC Challenge, inspired by ARC, provides controlled program synthesis tasks over synthetic images to evaluate automatic program construction, focusing on sequence, selection, and iteration. This set of 600 tasks has resisted automated solutions. This paper presents a structured inductive programming approach with LLMs that successfully solves tasks across all IPARC categories. The controlled nature of IPARC reveals insights into LLM-based code generation, including the importance of prior structuring, LLMs' ability to aid structuring (requiring human refinement), the need to freeze correct code, the efficiency of code reuse, and how LLM-generated code can spark human creativity. These findings suggest valuable mechanisms for human-LLM collaboration in tackling complex program synthesis.

Large Language Models (LLMs)' search capabilities have garnered significant attention. Existing benchmarks, such as OpenAI's BrowseComp, primarily focus on general search scenarios and fail to adequately address the specific demands of academic search. These demands include deeper literature tracing and organization, professional support for academic databases, the ability to navigate long-tail academic knowledge, and ensuring academic rigor. Here, we proposed ScholarSearch, the first dataset specifically designed to evaluate the complex information retrieval capabilities of Large Language Models (LLMs) in academic research. ScholarSearch possesses the following key characteristics: Academic Practicality, where question content closely mirrors real academic learning and research environments, avoiding deliberately misleading models; High Difficulty, with answers that are challenging for single models (e.g., Grok DeepSearch or Gemini Deep Research) to provide directly, often requiring at least three deep searches to derive; Concise Evaluation, where limiting conditions ensure answers are as unique as possible, accompanied by clear sources and brief solution explanations, greatly facilitating subsequent audit and verification, surpassing the current lack of analyzed search datasets both domestically and internationally; and Broad Coverage, as the dataset spans at least 15 different academic disciplines. Through ScholarSearch, we expect to more precisely measure and promote the performance improvement of LLMs in complex academic information retrieval tasks.

The rapid evolution of large language models (LLMs), driven by growing parameter scales, adoption of mixture-of-experts (MoE) architectures, and expanding context lengths, imposes unprecedented demands on AI infrastructure. Traditional AI clusters face limitations in compute intensity, memory bandwidth, inter-chip communication, and latency, compounded by variable workloads and strict service-level objectives. Addressing these issues requires fundamentally redesigned hardware-software integration. This paper introduces Huawei CloudMatrix, a next-generation AI datacenter architecture, realized in the production-grade CloudMatrix384 supernode. It integrates 384 Ascend 910 NPUs and 192 Kunpeng CPUs interconnected via an ultra-high-bandwidth Unified Bus (UB) network, enabling direct all-to-all communication and dynamic pooling of resources. These features optimize performance for communication-intensive operations, such as large-scale MoE expert parallelism and distributed key-value cache access. To fully leverage CloudMatrix384, we propose CloudMatrix-Infer, an advanced LLM serving solution incorporating three core innovations: a peer-to-peer serving architecture that independently scales prefill, decode, and caching; a large-scale expert parallelism strategy supporting EP320 via efficient UB-based token dispatch; and hardware-aware optimizations including specialized operators, microbatch-based pipelining, and INT8 quantization. Evaluation with the DeepSeek-R1 model shows CloudMatrix-Infer achieves state-of-the-art efficiency: prefill throughput of 6,688 tokens/s per NPU and decode throughput of 1,943 tokens/s per NPU (<50 ms TPOT). It effectively balances throughput and latency, sustaining 538 tokens/s per NPU even under stringent 15 ms latency constraints, while INT8 quantization maintains model accuracy across benchmarks.

Medical Question-Answering (QA) encompasses a broad spectrum of tasks, including multiple choice questions (MCQ), open-ended text generation, and complex computational reasoning. Despite this variety, a unified framework for delivering high-quality medical QA has yet to emerge. Although recent progress in reasoning-augmented large language models (LLMs) has shown promise, their ability to achieve comprehensive medical understanding is still largely unexplored. In this paper, we present Med-U1, a unified framework for robust reasoning across medical QA tasks with diverse output formats, ranging from MCQs to complex generation and computation tasks. Med-U1 employs pure large-scale reinforcement learning with mixed rule-based binary reward functions, incorporating a length penalty to manage output verbosity. With multi-objective reward optimization, Med-U1 directs LLMs to produce concise and verifiable reasoning chains. Empirical results reveal that Med-U1 significantly improves performance across multiple challenging Med-QA benchmarks, surpassing even larger specialized and proprietary models. Furthermore, Med-U1 demonstrates robust generalization to out-of-distribution (OOD) tasks. Extensive analysis presents insights into training strategies, reasoning chain length control, and reward design for medical LLMs. Our code is available here.