搜索结果：Opus

This paper introduces the Opus Workflow Evaluation Framework, a probabilistic-normative formulation for quantifying Workflow quality and efficiency. It integrates notions of correctness, reliability, and cost into a coherent mathematical model that enables direct comparison, scoring, and optimization of Workflows. The framework combines the Opus Workflow Reward, a probabilistic function estimating expected performance through success likelihood, resource usage, and output gain, with the Opus Workflow Normative Penalties, a set of measurable functions capturing structural and informational quality across Cohesion, Coupling, Observability, and Information Hygiene. It supports automated Workflow assessment, ranking, and optimization within modern automation systems such as Opus and can be integrated into Reinforcement Learning loops to guide Workflow discovery and refinement. In this paper, we introduce the Opus Workflow Reward model that formalizes Workflow success as a probabilistic expectation over costs and outcomes. We define measurable Opus Workflow Normative Penalties capturing structural, semantic, and signal-related properties of Workflows. Finally, we propose a unified optim

We evaluate the adversarial robustness of two frontier large language models (LLMs) developed by Anthropic, Fable 5 and Opus 4.8, against four families of automated jailbreak attack across 7 826 harmful intents spanning a ten-category harm taxonomy. Using the HackAgent red-teaming framework, hundreds of thousands of adversarial attempts were generated and every apparent success was independently re-adjudicated by a panel of three judge models (majority vote). Both models resist the majority of attacks, but the residual surface is larger than aggregate framing suggests: it is dominated by adaptive iterative attacks, while static obfuscation is near-fully neutralised. The strongest adaptive search (tree-of-attacks) breaks Opus 4.8 on 11.5% of intents overall, whereas Fable 5 stays in the single digits (6.1% worst-case). Aggregate rates therefore should not be read as reassurance. Even in these hardened configurations, the two models produced 1 620 (Opus 4.8) and 702 (Fable 5) panel-confirmed harmful completions spanning every harm category, located automatically, cheaply, and within the first one or two refinement steps by an attacker model with no human expert in the loop. The reaso

Artificial intelligence (AI) control protocols assume that trusted large language model (LLM) monitors reliably assess proposed actions across all deployment contexts. This paper tests that assumption in the geographic dimension. We audit Claude Opus 4.6-the monitor specified in Apart Research's AI Control Hackathon Track 3 benchmark-for systematic gaps in its factual knowledge of the global AI landscape. We develop the AI Control Knowledge Framework (ACKF), a six-dimension thematic scheme, and operationalise it with 17 verified indicators drawn from the Global AI Dataset v2 (GAID v2): 24,453 indicators across 227 countries published on Harvard Dataverse. A five-category response classification scheme distinguishes verifiable fabrication (VF) from honest refusal (HR); logistic regression with country-clustered standard errors combined with difference-in-differences (DiD) estimation quantifies geographic disparities in monitor accuracy across 2,820 country-metric-year observations. Contrary to our initial hypothesis, Claude Opus 4.6 produces higher fabrication rates for Global North queries than for Global South counterparts-a pattern consistent with a partial-knowledge mechanism in

This paper introduces the Opus Prompt Intention Framework, designed to improve complex Workflow Generation with instruction-tuned Large Language Models (LLMs). We propose an intermediate Intention Capture layer between user queries and Workflow Generation, implementing the Opus Workflow Intention Framework, which consists of extracting Workflow Signals from user queries, interpreting them into structured Workflow Intention objects, and generating Workflows based on these Intentions. Our results show that this layer enables LLMs to produce logical and meaningful outputs that scale reliably as query complexity increases. On a synthetic benchmark of 1,000 multi-intent query-Workflow(s) pairs, applying the Opus Prompt Intention Framework to Workflow Generation yields consistent improvements in semantic Workflow similarity metrics. In this paper, we introduce the Opus Prompt Intention Framework by applying the concepts of Workflow Signal and Workflow Intention to LLM-driven Workflow Generation. We present a reproducible, customizable LLM-based Intention Capture system to extract Workflow Signals and Workflow Intentions from user queries. Finally, we provide empirical evidence that the p

Rail-optimized network fabrics have become the de facto datacenter scale-out fabric for large-scale ML training. However, the use of high-radix electrical switches to provide all-to-all connectivity in rails imposes substantial power and cost. We propose a rethinking of the rail abstraction by retaining its communication semantics, but realizing it using optical circuit switches. The key challenge is that optical switches support one-to-one connectivity at a time, limiting the fan-out of traffic in ML workloads using hybrid parallelisms. We overcome this through \emph{parallelism-driven rail reconfiguration}, which exploits the non-overlapping communication phases of different parallelism dimensions. This time-multiplexes a single set of physical ports across circuit configurations tailored to each phase within a training iteration. We design and implement Opus, a control plane that orchestrates this in-job reconfiguration of photonic rails at parallelism phase boundaries, and evaluate it on a physical OCS testbed, the Perlmutter supercomputer, and in simulation at up to 2,048 GPUs. Our results show that photonic rails can achieve over $23\times$ network power reduction and $4\time

We report on an experiment in which Claude Opus~4.6, equipped with a suite of Model Context Protocol (MCP) tools for the Rocq proof assistant, autonomously proved 10 of 12 problems from the 2025 Putnam Mathematical Competition. The MCP tools, designed with Claude by analyzing logs from a prior experiment on miniF2F-Rocq, encode a "compile-first, interactive-fallback" strategy. Running on an isolated VM with no internet access, the agent deployed 141 subagents over 17.7 hours of active compute (51.6h wall-clock), consuming approximately 1.9 billion tokens. All proofs are publicly available.

When an LLM deobfuscates JavaScript, can poisoned identifier names in the string table survive into the model's reconstructed code, even when the model demonstrably understands the correct semantics? Using Claude Opus 4.6 across 192 inference runs on two code archetypes (force-directed graph simulation, A* pathfinding; 50 conditions, N=3-6), we found three consistent patterns: (1) Poisoned names persisted in every baseline run on both artifacts (physics: 8/8; pathfinding: 5/5). Matched controls showed this extends to terms with zero semantic fit when the string table does not form a coherent alternative domain. (2) Persistence coexisted with correct semantic commentary: in 15/17 runs the model wrote wrong variable names while correctly describing the actual operation in comments. (3) Task framing changed persistence: explicit verification prompts had no effect (12/12 across 4 variants), but reframing from "deobfuscate this" to "write a fresh implementation" reduced propagation from 100% to 0-20% on physics and to 0% on pathfinding, while preserving the checked algorithmic structure. Matched-control experiments showed zero-fit terms persist at the same rate when the replacement tabl

Trusted Execution Environments (TEEs) (e.g., Intel SGX and ArmTrustZone) aim to protect sensitive computation from a compromised operating system, yet real deployments remain vulnerable to microarchitectural leakage, side-channel attacks, and fault injection. In parallel, security teams increasingly rely on Large Language Model (LLM) assistants as security advisors for TEE architecture review, mitigation planning, and vulnerability triage. This creates a socio-technical risk surface: assistants may hallucinate TEE mechanisms, overclaim guarantees (e.g., what attestation does and does not establish), or behave unsafely under adversarial prompting. We present a red-teaming study of two prevalently deployed LLM assistants in the role of TEE security advisors: ChatGPT-5.2 and Claude Opus-4.6, focusing on the inherent limitations and transferability of prompt-induced failures across LLMs. We introduce TEE-RedBench, a TEE-grounded evaluation methodology comprising (i) a TEE-specific threat model for LLM-mediated security work, (ii) a structured prompt suite spanning SGX and TrustZone architecture, attestation and key management, threat modeling, and non-operational mitigation guidance, a

We present Yukthi Opus (YO), a multi-chain hybrid metaheuristic designed for NP-hard optimization under explicit evaluation budget constraints. YO integrates three complementary mechanisms in a structured two-phase architecture: Markov Chain Monte Carlo (MCMC) for global exploration, greedy local search for exploitation, and simulated annealing with adaptive reheating to enable controlled escape from local minima. A dedicated burn-in phase allocates evaluations to probabilistic exploration, after which a hybrid optimization loop refines promising candidates. YO further incorporates a spatial blacklist mechanism to avoid repeated evaluation of poor regions and a multi-chain execution strategy to improve robustness and reduce sensitivity to initialization. We evaluate YO on three benchmarks: the Rastrigin function (5D) with ablation studies, the Traveling Salesman Problem with 50 to 200 cities, and the Rosenbrock function (5D) with comparisons against established optimizers including CMA-ES, Bayesian optimization, and accelerated particle swarm optimization. Results show that MCMC exploration and greedy refinement are critical for solution quality, while simulated annealing and multi

Vertical Federated Learning (VFL) enables organizations with disjoint feature spaces but shared user bases to collaboratively train models without sharing raw data. However, existing VFL systems face critical limitations: they often lack effective incentive mechanisms, struggle to balance privacy-utility tradeoffs, and fail to accommodate clients with heterogeneous resource capabilities. These challenges hinder meaningful participation, degrade model performance, and limit practical deployment. To address these issues, we propose OPUS-VFL, an Optimal Privacy-Utility tradeoff Strategy for VFL. OPUS-VFL introduces a novel, privacy-aware incentive mechanism that rewards clients based on a principled combination of model contribution, privacy preservation, and resource investment. It employs a lightweight leave-one-out (LOO) strategy to quantify feature importance per client, and integrates an adaptive differential privacy mechanism that enables clients to dynamically calibrate noise levels to optimize their individual utility. Our framework is designed to be scalable, budget-balanced, and robust to inference and poisoning attacks. Extensive experiments on benchmark datasets (MNIST, CI

The advent of Generative Artificial Intelligence (GenAI) has brought a significant change to our society. GenAI can be applied across numerous fields, with particular relevance in cybersecurity. Among the various areas of application, its use in penetration testing (pentesting) or ethical hacking processes is of special interest. In this paper, we have analyzed the potential of leading generic-purpose GenAI tools-Claude Opus, GPT-4 from ChatGPT, and Copilot-in augmenting the penetration testing process as defined by the Penetration Testing Execution Standard (PTES). Our analysis involved evaluating each tool across all PTES phases within a controlled virtualized environment. The findings reveal that, while these tools cannot fully automate the pentesting process, they provide substantial support by enhancing efficiency and effectiveness in specific tasks. Notably, all tools demonstrated utility; however, Claude Opus consistently outperformed the others in our experimental scenarios.

In this paper, we explore the capabilities of state-of-the-art large language models (LLMs) such as GPT-4, Claude 3 Opus, and Gemini 1.0 Ultra in solving undergraduate-level control problems. Controls provides an interesting case study for LLM reasoning due to its combination of mathematical theory and engineering design. We introduce ControlBench, a benchmark dataset tailored to reflect the breadth, depth, and complexity of classical control design. We use this dataset to study and evaluate the problem-solving abilities of these LLMs in the context of control engineering. We present evaluations conducted by a panel of human experts, providing insights into the accuracy, reasoning, and explanatory prowess of LLMs in control engineering. Our analysis reveals the strengths and limitations of each LLM in the context of classical control, and our results imply that Claude 3 Opus has become the state-of-the-art LLM for solving undergraduate control problems. Our study serves as an initial step towards the broader goal of employing artificial general intelligence in control engineering.

This paper introduces Opus, a novel framework for generating and optimizing Workflows tailored to complex Business Process Outsourcing (BPO) use cases, focusing on cost reduction and quality enhancement while adhering to established industry processes and operational constraints. Our approach generates executable Workflows from Intention, defined as the alignment of Client Input, Client Output, and Process Context. These Workflows are represented as Directed Acyclic Graphs (DAGs), with nodes as Tasks consisting of sequences of executable Instructions, including tools and human expert reviews. We adopt a two-phase methodology: Workflow Generation and Workflow Optimization. In the Generation phase, Workflows are generated using a Large Work Model (LWM) informed by a Work Knowledge Graph (WKG) that encodes domain-specific procedural and operational knowledge. In the Optimization phase, Workflows are transformed into Workflow Graphs (WFGs), where optimal Workflows are determined through path optimization. Our experiments demonstrate that state-of-the-art Large Language Models (LLMs) face challenges in reliably retrieving detailed process data as well as generating industry-compliant wo