All News

1539 technical news articles

Canary Tools Divides Agent Tool-Misselection into Six Categories, With an Approximately 36-Fold Difference in Deception Rates Across Eight Models

A new benchmark plants targeted probes—including semantic lures, parameter traps, and capability illusions—in MCP tool lists, using erroneous calls to pinpoint agents’ tool-selection weaknesses. Across 11,520 experiments, model tier did not consistently predict safety, while the ability to recover after being deceived raised task success rates from 16% to 52%.

2026-08-06

代理工具

Argus Enables Fixed-Weight Agents to Accumulate Experience, Raising SWE-Bench Pro Accuracy From 59% to About 78%

Argus breaks long-running tasks into constrained missions, with a Manager, Planner, Engineer, and Reviewer jointly maintaining persistent state. Its authors report improving program repair performance at 1.41× the token cost; an agent-generated RWKV6 kernel was also merged into an open-source project after external corrections.

2026-08-06

推論系統

ReCo Coordinates KV Cache and Reasoning Length with Step-Level Rewards, Cutting End-to-End Latency by More Than Half

ReCo dynamically compresses the KV cache based on process rewards for each reasoning step, while suppressing repetitive reflection and enabling early stopping. In the authors’ experiments across three reasoning models and six benchmarks, generated tokens fell by 37% to 65%, while latency improved by 2.08× to 2.35×.

2026-08-06

代理工具

ContextWeave Re-Evaluates Agent Memory with 1,000 Executable Office Tasks, Raising Preference Score from 41.50 to 70.60

ContextWeave goes beyond asking whether agents can retrieve old information, examining whether memory improves subsequent document and office work. Experiments show that retaining complete, actionable experiences outperforms short summaries, although incorrect recollections are also more likely to affect execution results.

2026-08-06

代理工具

MatrAIx Synthesizes 8.3 Billion Personas Using 1,290-Dimensional Attributes, Open-Sources Million-Record Core Dataset for AI Product Testing

MatrAIx combines correlated demographic attributes to create 8.3 billion simulated users and lets persona agents interact with surveys, chatbots, websites, and applications. The team has released a core dataset of roughly one million records and its evaluation code, but current validation shows only that the agents follow assigned personas—not that they can replace testing with real people.

2026-08-06

AI 基礎設施

Azure Tests Reveal CPU–GPU Fragmentation in Agentic Workflows, While Agora Reclaims Idle Compute Without Sacrificing Tail Latency

Microsoft Azure research shows that agents repeatedly switch among model inference, tools, and orchestrators, putting the CPU back on the critical path while GPUs often sit idle. The Agora prototype can reclaim compute based on workflow patterns, but incorrectly reducing GPU allocations for parallel agents can worsen tail latency by 16×.

2026-08-06

OctoLong Trains Long-Context Models on Cross-Library Dependency Chains, Lifting the 8B Model’s RepoQA Score to 64.63

Instead of simply concatenating individual libraries into long documents, OctoLong recursively gathers implementations by following AST, language-server, and package dependency relationships. The research team released models ranging from 600M to 14B parameters, with experiments showing that dependency-dense data improves code retrieval, state tracking, and API usage.

2026-08-06

代理工具

ABSeeker Backtracks from Correct Answers to Derive Search Clues, Raising a 4B Model’s BrowseComp Score to 55.3%

ABSeeker breaks down a search agent’s final success or failure into step-level rewards, preserving useful queries from failed trajectories while downweighting erroneous steps in successful ones. The team has released the 4B model weights and the inference and training pipelines, though its best results still rely on additional context management and an LLM judge.

2026-08-06

代理工具

KV Cache Is Not “Latent Thought”: Mismatched Caches Reveal That Multi-Agent Gains May Be Merely Interface Effects

A new study re-audits latent communication in multi-agent systems using mismatched, zeroed, and random KV caches, finding that the large performance drop caused by removing the cache does not necessarily prove that the receiver used specific information from the sender. On some standard benchmarks, zeroing the cache reduced accuracy by 14.7 points, while replacing it with a cache from another problem cost only 0.4 points.

2026-08-06

AI 安全

Small Models’ Self-Reported Confidence Cannot Directly Control Risk; Only 3 of 22 Settings Meet the 20% Error Cap

A new study tests whether “model-reported confidence” can determine when to hand a task off to a human across 11 open-weight models. Although Platt scaling reduces ECE to as low as 0.02, rigorous finite-sample validation rejects every autonomous-answering setting under a 10% risk cap.

2026-08-06

WorldCycle Creates Its Own Verification Signals with Reversible Actions, Quadrupling Video World Model Accuracy on Compositional Actions

WorldCycle combines actions and their inverse operations into closed loops, enabling long-term state drift to be measured without ground-truth future videos. In experiments with an 8B video world model, it reduced cycle errors by up to 44% and increased compositional-action accuracy from 0.136 to 0.553.

2026-08-06

AI 安全

New Benchmark Finds Implicit Bias Can Evade Chain-of-Thought Monitoring, With Detection Rates as Low as 5% in Some Settings

A new study uses paired prompts to compare whether models leave detectable evidence in their chains of thought when influenced by explicit instructions versus implicit cues. Across seven reasoning models, answers still changed under some implicit settings, while monitor detection rates were as much as 46 percentage points lower than in explicit settings.

2026-08-06