Pai 的 AI 雷达

今日必看

含近几日仍热话题，按时间衰减排序。今天先看 Tencent/WeKnora，其次 langbot-app/LangBot。并行还有 Enterprise-managed sandbox in Copilot for JetBrains、How GPT‑5.6 Sol helps run quantum computing experiments。

日期
2026/09/09
今日条目
73
往期2026/09/09
全部73编码代理0Agent与工具2个人助理2模型与评测24创作成片9开发与开源16产品接入20
领域机器人具身8生信医疗2法律合规2金融1安全治理1行业观察10

更多资讯

#23

D3ARC: Time-Critical Distributed Disaster Detection for Asynchronous Cooperative Multi-Robot Systems

arXiv:2609.07350v1 Announce Type: cross Abstract: Climate change is increasing the severity and unpredictability of natural disasters. In time-critical crises such as wildfires, traditional monitoring practices remain limited by coverage, cost, and personnel risk, paving the way

推荐理由D3ARC: Time-Critical Distributed：arXiv:2609.07350v1 Announce Type: cross Abstract: Climate change is increasing。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#24

GIF: Agentic Generation of Interactive and Functional Object Compositions for Robot Learning

arXiv:2609.05927v1 Announce Type: new Abstract: Robot manipulation foundation models require scalable evaluation and data generation across diverse scenarios, with simulation providing an environment for both. Automated scene generation offers a promising path, yet prior work has

推荐理由GIF: Agentic Generation of Interactive：arXiv:2609.05927v1 Announce Type: new Abstract: Robot manipulation foundation models。
模型与评测arXiv cs.RO 官方
arxiv.org打开原文
#25

$A^2E$ : An End-to-End Agent Auditing Engine

arXiv:2608.07346v4 Announce Type: replace Abstract: With the rapid advancement of large language models (LLMs), harnesses have become essential infrastructure for deploying agents across a wide range of domains. The fast-evolving harness ecosystem has also made rigorous capabilit

推荐理由$A^2E$ : An End-to-End Agent Auditing：arXiv:2608.07346v4 Announce Type: replace Abstract: With the rapid advancement of large。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#26

A Three-Tier Persona Vector for Controllable User Simulation in Agentic Evaluation

arXiv:2609.08592v1 Announce Type: new Abstract: Evaluating tool-augmented LLM agents requires diverse, realistic user inputs yet most evaluation frameworks use flat role descriptions ("you are an angry customer") that produce near-identical conversations regardless of the underly

推荐理由A Three-Tier Persona Vector：arXiv:2609.08592v1 Announce Type: new Abstract: Evaluating tool-augmented LLM agents。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#27

A visual large language foundational model for medical image recognition using clinician-oriented social media

arXiv:2609.06914v1 Announce Type: new Abstract: Large language models (LLMs) have demonstrated strong capabilities across diverse domains, showing considerable potential in medicine. However, their application in medical settings remains limited by the scarcity of visual question

推荐理由A visual large language foundational：arXiv:2609.06914v1 Announce Type: new Abstract: Large language models (LLMs) have。
模型与评测生信医疗arXiv cs.AI 官方
arxiv.org打开原文
#28

AgentBrew: Offline Tool-Use Agent Learning from Raw Real-World Trajectories

arXiv:2609.05837v1 Announce Type: new Abstract: LLM-based agents are increasingly deployed in real-world applications through tool-use APIs, yet training them for specific environments remains fundamentally difficult: real-world applications provide no pre-defined tasks or verifi

推荐理由AgentBrew: Offline Tool-Use Agent：arXiv:2609.05837v1 Announce Type: new Abstract: LLM-based agents are increasingly。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#29

Agentic BAIM-LLM Evaluation (ABLE): Benchmarking LLM Use of Protein Design Tools

arXiv:2609.05818v1 Announce Type: new Abstract: We introduce ABLE, a benchmark for evaluating LLM agents' ability to use biological AI models (BAIMs), such as ProteinMPNN and AlphaFold3, in dual-use protein design workflows. ABLE assesses agent performance through a set of tasks

推荐理由Agentic BAIM-LLM Evaluation (ABLE)：arXiv:2609.05818v1 Announce Type: new Abstract: We introduce ABLE, a benchmark。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#30

Agentic Electronic Design Automation: A Handoff Perspective

arXiv:2606.19795v2 Announce Type: replace-cross Abstract: Electronic design automation (EDA) is multi-stage and handoff-heavy, relying on transfers among humans, design artifacts, and multiple tools. LLM-based agents now participate in these transfers, yet the resulting research

推荐理由Agentic Electronic Design Automation:：arXiv:2606.19795v2 Announce Type: replace-cross Abstract: Electronic design automation。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#31

AnchorScore: A CLIP-Based Diagnostic of MLLM Annotation Difficulty

arXiv:2608.16690v2 Announce Type: replace Abstract: Multimodal large language models (MLLMs) are widely used for automated annotation, yet their per-class accuracy varies widely (e.g., 12%-98% across the 13 classes of three classroom sub-datasets) and is expensive to measure: eva

推荐理由AnchorScore: A CLIP-Based Diagnostic：arXiv:2608.16690v2 Announce Type: replace Abstract: Multimodal large language models。
模型与评测arXiv cs.CV 官方
arxiv.org打开原文
#32

AutoLexSteer: Automatic Contrast Construction for Lexical Activation Steering

arXiv:2609.06879v1 Announce Type: cross Abstract: Steering vectors have rapidly emerged as a popular and effective method for guiding the output of LLMs in very specific ways. But constructing accurate steering vectors is a difficult manual process due to the opacity of embedding

推荐理由AutoLexSteer: Automatic Contrast：arXiv:2609.06879v1 Announce Type: cross Abstract: Steering vectors have rapidly emerged。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#33

Bait-and-Recover: Poisoning Internal Refusal Signals to Defend LLMs against White-Box Editing Jailbreaks

arXiv:2609.05794v1 Announce Type: cross Abstract: Open-weight large language models face a low-cost white-box threat from representation engineering attacks. Attackers can estimate refusal directions and search for projection-matrix edits that suppress safety alignment while pres

推荐理由Bait-and-Recover: Poisoning Internal：arXiv:2609.05794v1 Announce Type: cross Abstract: Open-weight large language models。
模型与评测安全治理arXiv cs.AI 官方
arxiv.org打开原文
#34

Beyond Prompts: Measuring and Optimizing LLM Tool-Agent Harnesses

arXiv:2609.05736v1 Announce Type: new Abstract: LLM tool agents can be improved without retraining by modifying the runtime harness around a fixed model: prompts, tool interfaces, middleware, state handling, and recovery logic. We study this setting as resource-bounded harness se

推荐理由Beyond Prompts: Measuring：arXiv:2609.05736v1 Announce Type: new Abstract: LLM tool agents can be improved without。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#35

Competing at Every Price Point with Agentic Evolution over a Menu of LLMs

arXiv:2608.16207v2 Announce Type: replace Abstract: Consider a firm that surveys its competition for a particular agentic task and seeks to offer superior accuracy at every price point. A firm that Pareto-dominated its competitors would leave no rational customer a reason to buy

推荐理由Competing at Every Price Point：arXiv:2608.16207v2 Announce Type: replace Abstract: Consider a firm that surveys its。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#36

Cost-Optimal LLM Routing with Limited User Feedback under User Satisfaction Guarantees

arXiv:2606.19376v2 Announce Type: replace-cross Abstract: Inference costs for large language model (LLM) applications are rapidly growing, driven by surging demand and rising infrastructure cost. Users expect high-quality responses, and in commercial settings this is formally cod

推荐理由Cost-Optimal LLM Routing with Limited：arXiv:2606.19376v2 Announce Type: replace-cross Abstract: Inference costs for large。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#37

Elastic Horizon: Discovering the Effective Interaction Frontier in Agentic Reinforcement Learning

arXiv:2609.07247v1 Announce Type: new Abstract: Scaling the interaction horizon-the maximum number of environment interactions per episode-improves LLM agents on long-horizon tasks, and curriculum-based methods that progressively expand the horizon outperform fixed-horizon altern

推荐理由Elastic Horizon: Discovering：arXiv:2609.07247v1 Announce Type: new Abstract: Scaling the interaction horizon-。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#38

EnvCraft: Synthesizing Executable Environments in Agentic RL for Claw-like Agent

arXiv:2609.05576v1 Announce Type: new Abstract: The paradigm of LLMs has rapidly shifted from passive language interfaces to autonomous Claw-like agents that execute long-horizon tasks across stateful workspaces. While Agentic Reinforcement Learning (Agentic RL) provides a promis

推荐理由EnvCraft: Synthesizing Executable：arXiv:2609.05576v1 Announce Type: new Abstract: The paradigm of LLMs has rapidly。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#39

FedSubMuon: Communication-Efficient Federated LLM Fine-Tuning via Structured Subspace Muon

arXiv:2609.06073v1 Announce Type: cross Abstract: Federated fine-tuning adapts large language models (LLMs) to decentralized client data, but its scalability in cross-device training is often limited by the high communication cost. Muon is an optimizer that improves optimization

推荐理由FedSubMuon: Communication-Efficient：arXiv:2609.06073v1 Announce Type: cross Abstract: Federated fine-tuning adapts large。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#40

From LLM-Generated Specifications to Learned Quadruped Locomotion

arXiv:2609.07111v1 Announce Type: cross Abstract: Quadruped robot locomotion policies are often trained using reinforcement learning, which in turn relies heavily on hand-crafted reward functions. Designing reward functions requires substantial manual engineering, and it is often

推荐理由From LLM-Generated Specifications：arXiv:2609.07111v1 Announce Type: cross Abstract: Quadruped robot locomotion policies。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#41

Inference-Time Graph Engineering for Multi-Agent LLM Workflows

arXiv:2609.05774v1 Announce Type: new Abstract: Recent multi-agent LLM systems increasingly rely on graph-structured communication to coordinate specialized agents. We revisit multi-agent orchestration from a graph-engineering perspective: rather than optimizing a static topology

推荐理由Inference-Time Graph Engineering：arXiv:2609.05774v1 Announce Type: new Abstract: Recent multi-agent LLM systems。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#42

LLMs for Social Network Modeling: From Network Generation to Dynamic Processes

arXiv:2609.08049v1 Announce Type: cross Abstract: Large language models (LLMs) are rapidly emerging as a new paradigm for modeling social networks by representing users and their relationships and interactions through natural language. Unlike classical network models or deep lear

推荐理由LLMs for Social Network Modeling: From：arXiv:2609.08049v1 Announce Type: cross Abstract: Large language models (LLMs) are。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#43

Localizing Emergent Failures in Agentic AI: Recovering Minimal Repair Families via Counterfactual Replay

arXiv:2608.29228v2 Announce Type: replace Abstract: Failures in agentic AI systems can arise from interactions among messages exchanged by multiple large language model (LLM) agents. Pointwise attribution cannot distinguish a jointly necessary repair from alternative singleton re

推荐理由Localizing Emergent Failures in Agentic：arXiv:2608.29228v2 Announce Type: replace Abstract: Failures in agentic AI systems can。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文
#44

MV-STRIDE: Enabling MLLMs to Master Multi-View Spatial Reasoning via Hierarchical Capability Modeling

arXiv:2609.07258v1 Announce Type: cross Abstract: Despite the rapid progress of Multimodal Large Language Models (MLLMs) in 2D vision-language tasks, robust multi-view spatial reasoning remains a fundamental bottleneck due to the lack of structured 3D cognitive pathways in existi

推荐理由MV-STRIDE: Enabling MLLMs to Master：arXiv:2609.07258v1 Announce Type: cross Abstract: Despite the rapid progress。
模型与评测arXiv cs.AI 官方
arxiv.org打开原文