California Gov. Gavin Newsom (D) is positioning the state to take the lead on AI oversight, including the potential to mandate a "kill switch" for…
Deploy production-ready Hugging Face models on Amazon SageMaker AI using six open-source agent skills. Point a coding agent at a model and get back a real-time endpoint with the right serving container, autoscaling, Amazon CloudWatch alarms, and a verified teardown path.
Assine o Pro, sem anúncios
Faça upgrade para uma leitura sem interrupções e acesso prioritário a novas fontes.
Ver preçosAs scaling pre-training data alone yields diminishing returns, post-training is becoming increasingly important across physical AI domains such as autonomous driving. End-to-end driving policies are pre-trained in open loop with behavior cloning on human demonstrations. However, compounding errors during closed-loop deployment can take the vehicle outside the training data distribution, increasing the risk of safety-critical incidents. Closed-loop post-training can mitigate this risk but requires costly simulation for sensor-based policies. We propose OPTED (on-policy fine-tuning for end-to-end driving) which decouples reinforcement learning from the post-training of the end-to-end policy: a privileged teacher is trained using RL on vectorized inputs (HD-map and bounding boxes). This teacher then provides supervision to the pre-trained student during closed-loop post-training. We apply OPTED to two camera-based models, TransFuser and VaVAM, and fine-tune them in AlpaSim, using neural reconstructions (3DGS) of real driving logs. Driving scores increase by factors of 1.6and 9.5, respectively. In controlled experiments OPTED matches closed-loop performance with approximately three orders of magnitude fewer simulator interactions than direct RL post-training, while staying closer to the human prior. Project page: https://01dami23.github.io/opted/
Base Labs, the research group Baseten spun up earlier this year, will develop and publish methods for training and monitoring open models.
Large language models are increasingly used in healthcare communication, yet most evaluations emphasize response quality while assuming that the user's concern has been interpreted correctly. We introduce HerHealthEval, a controlled evaluation framework for multilingual understanding of women's-health communication. For each clinical case, HerHealthEval provides matched versions in English, French, and Modern Standard Arabic using six communicative forms: canonical, clinical, layperson, indirect or hedged, emotionally concerned, and deliberately under-specified. The first five express the same underlying concern and retain the same clinical information, whereas the under-specified form intentionally omits relevant details to test whether the model recognizes that clarification is needed. We evaluate a multilingual instruction model and QLoRA-adapted variants on concern classification, risk calibration, clarification behavior, parse compliance, and cross-form consistency. Results reveal that aggregate accuracy and consistency can conceal safety-relevant failures. A multilingual adaptation model reaches 0.994 under-triage in French and Arabic under language-asymmetric risk supervision. A controlled re-adaptation using source-derived, language-invariant risk labels reduces under-triage to 0.572 and 0.558, respectively. These findings show that robust multilingual healthcare evaluation requires explicit testing of register variation, uncertainty handling, and the provenance and invariance of adaptation labels.
Wood Mackenzie built APEX, a shared agentic AI platform on Amazon Bedrock AgentCore so every team can ship production agents without rebuilding runtime, identity, observability, and guardrails from scratch. Learn why they chose AgentCore, how APEX Studio operates it, and where multi-agent systems go next.
Traffic regulations are written for human interpretation and therefore rely on shared background knowledge and flexible phrasing, which inherently introduce ambiguity, context dependence, and semantic underspecification. These linguistic characteristics conflict with the precision required by computational reasoning engines such as Prolog, which demand explicit logical structure. This study evaluates two baseline translation approaches, Natural Language to Prolog (NLProlog) and Logical English to Prolog (LEProlog), and introduces a new reasoning-guided translation framework called Structured Four-Stage Legal Translation (S4LProlog). The proposed S4L framework performs semantic role extraction, scene completion, logical mapping, and Prolog rule generation within a single guided prompt, enabling direct translation of raw traffic rules into executable logic without human intervention. A benchmark consisting of twenty real-world traffic rules was used to evaluate each approach in terms of syntactic validity, semantic correctness, and logical completeness. S4LProlog achieves the highest accuracy, correctly formalizing 75 percent of the rules, while NLProlog reaches 60 percent and LEProlog reaches 55 percent. Qualitative analysis further shows that S4L captures implicit causal relations, deontic modality, and exception structure more reliably than the baselines. These results demonstrate that structured reasoning prompts can substantially improve the reliability of natural-language-to-logic translation for legal and safety-critical applications.
Subliminal learning allows language models to transmit behavioral traits through training data with no obvious semantic relationship to those traits, undermining content-based data filtering as a safety intervention. Training data attribution offers an alternative: it identifies the training examples responsible for a given model behavior, independent of their semantic content, and so may apply in exactly the cases where semantic inspection fails. We evaluate three gradient-based attribution methods (GradCos, a contrastive GradCos variant, and EK-FAC) across three models, comparing them against divergence tokens, a strong baseline previously shown to localize subliminal learning (albeit one that requires access to counterfactual teacher models). Filtering at the token level, EK-FAC mitigates a significant part of the effect, the other methods provide little benefit, and all mostly fall short of divergence tokens. Filtering entire samples is less effective for every method, though EK-FAC often gives a stronger signal than divergence tokens in this setting. Success is inconsistent across methods and settings: variants that work well for some model-preference combinations fail for others, and we do not identify a consistent explanation for these differences. Our results suggest that gradient-based attribution can identify data responsible for subliminal learning in some settings, but that some approximations are more reliable than others.
Emergent coordinated behaviors of AI agents are starting to present critical safety risks. A key phenomenon driving these behaviors is the rapid formation and spread of beliefs about the world, and mechanistic understanding is crucial for collective alignment. To this end, we introduce the Flag Game, a toy model for studying the mechanisms of collective belief formation. Concretely, a hidden country flag defines the ground truth, and each bounded agent directly observes only a private crop but can exchange beliefs and weigh social evidence from peers. Despite its simplicity, the Flag Game reproduces rich collective phenomenology: non-monotonic scaling of performance with population size, accuracy gains from social-awareness prompting and team diversity, and strong effects of organizational structure. In particular, we identify that collective belief collapse at small population sizes turns into collective belief polarization as the population grows. This polarization causes the performance decline at large population sizes, but creates diversity in collective beliefs. Finally, we dissect the mechanisms underlying collective belief collapse and polarization with two complementary approaches. We first introduce social circuit attribution, a technique to predict which agent, and what view, matters most to collective dynamics, and verify its predictions by causal interventions on agents, tracing how agent patching changes collective outcomes. However, the efficacy of causal interventions on agents decreases as the population grows. We therefore develop a statistical mechanical theory for larger populations and verify that it matches the empirical phase diagram. Together, these results take a first step toward mechanistic swarm interpretability, a science of how the properties of individual agents and their communication give rise to emergent collective behavior.
Warning: This paper contains examples of stereotypes and social bias. LLMs are increasingly used in interactive settings by the general public, making the evaluation of model behavior in multi-turn conversational scenarios important for safety, including stereotyping-related harms. However, existing multi-turn social bias evaluations often rely on pre-specified or template-based user inputs that do not adapt to model responses and typically assume a fixed dialogue length in advance. In this paper, we study social bias dynamics in response-conditioned multi-turn interactions using a controlled evaluation protocol that generates follow-up user queries from the evolving dialogue history and allows evaluation over variable numbers of turns. Experimental results show that LLMs exhibit social bias even in coherent, response-conditioned multi-turn interactions, revealing late-emerging bias, non-monotonic bias patterns, and bias re-emergence. These results motivate evaluations that extend beyond fixed-turn, pre-scripted protocols. Our findings highlight the importance of analyzing social bias as a turn-level dynamic phenomenon.
Poll data released Tuesday by the New York Times and Siena University confirms what we've already been seeing, and what politicians are responding to …
When OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, Google DeepMind cofounder Demis Hassabis, and SpaceX head Elon Musk loosely agreed over the weekend to…
Chain-of-thought (CoT) monitoring is a safety strategy where the reasoning of a large language model "actor" is inspected by a "monitor" (often another language model) for signs of unsafe planning, deception, or misalignment. We find that planting harmful but benign-sounding reasoning in the actor's context can steer it to perform adversarial actions while evading monitors, an attack we term "plan injection". We initially discover this attack in the multiple-choice question-answering monitorability setting proposed by Lanham et al. (2023), using the investigator-agent elicitation framework of Li et al. (2025). We generalize the attack and show that the discovered behavior scales to harder tasks (achieving 25-33% monitor evasion rates across different monitorability benchmarks) and larger models such as DeepSeek-R1. Across the settings we study, actor models not only follow injected plans but also paraphrase them as their own reasoning, without explicit attribution to the injections. Finally, we find cases where extra monitor resources cause harm - giving the monitor access to the injected plan drops detection by as much as 50% in the Bio-Math task and in a case study on monitor reasoning budget, we find transcripts where additional thinking tokens are spent rationalizing the injected plan rather than flagging it.
Recurrent GNNs iterate message passing to convergence, and their logical characterizations to date rely on multi-set aggregation, graded (counting) logics, and halting or acceptance conditions that cannot be verified from the network's parameters. We study recurrent GNNs with set-based aggregation and identify sufficient conditions checkable from the weights for networks to compile into formulas and formulas into networks. The main result is an effective, two-directional equivalence between a class of networks and the Boolean closure of reachability and safety properties, the fragment BΣ^{}_1 of the modal μ-calculus. The fragment is not an artifact: it is the exact expressive level of stabilization over finite vocabulary, which supports fixed points of a single polarity and Boolean combinations thereof, but not the composition of fixed points of opposite polarities. The correspondence needs no counting logic, no external halting signal, and no non-effective acceptance condition, yielding a verifiable path from weights to symbolic explanations for networks meeting the conditions.
Large language models (LLMs) often learn both desirable and undesirable properties during post-training. We study whether midtraining, an earlier training stage, can shape which of these properties later generalise. We introduce Inoculation Midtraining, a technique that teaches a base model that unsafe behaviour belongs to a designated context, as indicated by the neologism (a new token) introduced during midtraining, and then post-trains the model on unsafe data within that context. We then evaluate the model outside the context, with the neologism excluded from the system prompt. Across supervised fine-tuning and reinforcement learning post-training regimes, we find that Inoculation Midtraining can reduce misalignment while preserving the transfer of benign data properties (e.g., speaking in German or Shakespearean prose). However, our approach does not outperform standard Inoculation Prompting, is sensitive to training configuration, and produces a leaky boundary that nearby contextual cues can reactivate. These results show that inoculation with a learned association introduced via midtraining can shape selective generalisation. Still, more work is needed before this approach can become a load-bearing component in a developer's safety framework.
% !TEX root = ../main.tex People increasingly use large language models (LLMs) for mental health support, yet their safety in evolving, high-risk conversations remains poorly characterised. We developed K-Bench, a clinician-calibrated, protected benchmark evaluating 125 model configurations representing 33 base models from 14 providers across a fixed cohort of 200 multi-turn vignettes involving suicide, self-harm, domestic violence, substance misuse, and no-risk presentations. Synthetic patient conversations showed substantial distributional overlap with real human-AI conversations. A frozen GPT-4o judge achieved 94.2% exact agreement with clinician consensus across 6,751 eligible item comparisons from 151 clinician-rated transcripts. Leading models combined strong supportive conversation with combined-risk scores above 95, whereas risk exploration exposed substantial variation among lower-performing configurations. Therapeutic prompting produced configuration-specific gains concentrated among weaker models, while elevated reasoning produced no average improvement. K-Bench combines broader clinical coverage and configuration-scale comparison with a continuously updated public leaderboard whose operational test materials are protected from direct optimisation. The leaderboard is available at www.k-bench.ai.
The code of conduct lays out general principles that Microsoft AI models should uphold — supporting humans rather than replacing them, for instance, and accelerating…
Clinical AI systems are evaluated with instruments built for research settings (reference-based similarity metrics and expert rubric panels) that measure resemblance to an artifact rather than reduction of a burden. We introduce KnowBench, pioneered by Knowtex, whose unifying metric is Effort Reduction (ER): the proportion of system-generated clinical work product accepted by the responsible clinician under expert and safety review. ER is defined once and instantiated per task across the administrative workload clinical AI automates: visit notes, diagnosis and billing codes, orders, EHR chart summarization, patient after-visit summaries, and clinical decision support. In every instantiation the construction is identical: the clinician's review-and-attestation event is the ground truth, every accepted unit is work the system completed, and every correction is residual effort returned to the clinician. The primary contribution of this paper is the benchmark itself: the metric, its degenerate cases, and a reporting protocol under which ER claims are auditable and cross-system comparable. Alongside it we report an initial headline measurement from the documentation instantiation: over one million signed encounters across a production window exceeding six months and thirteen medical specialties, Knowtex's proprietary fine-tuned clinical foundation models operating inside a closed feedback architecture achieve an aggregate ER of 97.99%, with per-specialty aggregates spanning 96.8-98.9%. This release reports the protocol's checklist partially, and states which companion statistics are withheld; the benchmark is offered so that this figure, and every figure reported after it, can be held to the same standard.
Pick the right generative AI customization approach on AWS with an 8-step decision framework, from prompt engineering and RAG to fine-tuning, continued pre-training, and Amazon Nova Forge. Start simple and escalate only when you must.
