All News

1533 technical news articles

Anthropic Tests a Fine-Tuned “Lie Detector”: In-Domain AUROC Rises to 0.95, but Falls to Around 0.70 on a Different Type of Lie

The research team used eight open model families to generate roughly 200,000 examples of pressure-induced scenarios, then applied LoRA to train models to determine whether they themselves were lying. The detectors mainly learned surface-level patterns such as sycophancy and capability denial, performing only slightly better than prompting baselines on unseen lie types.

2026-08-25

ReWorld Uses a Pose-Indexed Landmark Bank to Fix the KV Budget, Allowing Interactive World Models to Return to Earlier Scenes After Long Journeys

ReWorld assigns short-term action control and long-term scene memory to different attention ranges, then preserves history in a fixed-capacity landmark bank. The authors demonstrate real-time streaming at 704×1280 and 64-second return trajectories, but only the paper and demo page are currently available; neither model weights nor code have been released.

2026-08-25

GitHub Repo

Prime Agent Open-Sources a Persistent RLM Harness That Keeps Agent Memory, Subagents, and Execution Environments Outside the Conversation

Prime Agent supports long-running coding agents with a persistent IPython environment, daemon sessions, and a versionable Continual Harness. Its authors report that ARC-AGI-3 RHAE Best@1 rose from 30% to 95.5%, but the score reflects the complete model-and-harness system and should not be treated as an isolated comparison of underlying model capabilities.

2026-08-25

代理工具

SWE-bench Science Releases 119 Scientific Code-Repair Tasks, With the Best Agent Still Below a 50% Full Pass Rate

The new benchmark uses private scientific tests to check units, numerical invariants, file semantics, and cross-module data flow, rather than merely whether code compiles. Claude Code paired with Opus 5 achieved an overall pass@1 of 47.90%, while experiments also showed that adding domain knowledge can sometimes anchor agents in the wrong direction.

2026-08-25

GitHub Repo

DFlash 2 Drafts Tokens in Parallel for Qwen3.8-27B, but vLLM Nightly Currently Has Loading and VRAM Errors

DFlash 2 improves block-diffusion drafting with a candidate-path selector and dynamic convolution, delivering 2.7× to 3.4× the throughput of autoregressive decoding for Qwen3.8-27B in single-request tests. The latest reproducible issues show that vLLM nightly instantiates the wrong decoder layer and temporarily allocates nearly 4.74 GiB of extra memory before sharing vocabulary weights, potentially preventing startup on 24 GB GPUs.

2026-08-25

硬體

MTIA 300 Integrates 12 RDMA NICs and Collective Communication Engines On-Chip, Avoiding PCIe and Compute-Core Contention

Meta has disclosed the system design of MTIA 300, which accelerates recommendation-model training with built-in NIC chiplets, near-memory reduction units, and compiled HCCL. The company says communication time across 40 accelerators was reduced to roughly one-quarter that of a reference GPU cluster, but provided no comparisons of cost, power consumption, or general-purpose LLM training.

2026-08-25

AI 基礎設施

AWS ParallelCluster 3.16 Adds Node Diagnostics, but the Upgrade Also Changes IAM, NFS, and the Container Stack

The new AMI includes `pcluster-diag`, which can run role-specific checks of Slurm, IMDS, directory services, and Lustre status and produce structured reports. The upgrade also ends support for Amazon Linux 2 and AWS Batch, while introducing permission, network egress, and runtime changes that require advance preparation.

2026-08-25

代理工具

DeepSeek Harness rc.2’s Split npm Tags: The Main Package Is Current, but Plugin Development May Pull in 0.0.x

The `latest` tag for DeepSeek Harness’s main package points to 0.1.1-rc.2, but several underlying packages still mark early versions as the latest. Installing the plugin API directly may trigger peer dependency conflicts or silently create a mismatched dependency tree.

2026-08-25

AI 基礎設施

SageMaker HyperPod Adds Managed Ray: Training, Inference, and Failure Recovery Share the KubeRay Interface

AWS has integrated Ray cluster creation, remote job submission, monitoring, and workspace connectivity into HyperPod on EKS. The interface retains KubeRay and standard Ray APIs, but deployment still requires several add-ons, and AWS has not published cost or performance comparisons.

2026-08-25

AI 研究

MIT’s η-Learning Generates 100-Year Events Without Extreme Samples, but It Is Not a Disaster Timing Forecaster

η-learning constrains generative models with extreme value statistics to synthesize high-resolution extreme scenarios from paired data that lack rare events. The method can support infrastructure stress testing, but it cannot predict when disasters will occur and still depends on trustworthy external statistics.

2026-08-25

代理工具

GitHub Agentic Workflows Adds Real-Time PR Steering, but v0.87.4 Cited in the Official Weekly Report Does Not Include It

The new feature lets a running agent read PR comments and adjust its work based on `steer` instructions, while requiring explicit read permissions. However, the official weekly report recommends trying it with the earlier v0.87.4 release; deployments pinned to that tag will not actually get the feature.

2026-08-25

AI 基礎設施

NVIDIA Groq 3 LPX Enters Mass Production: Static Scheduling Delivers 3,431 Output Tokens per Second at 100K Context

Groq 3 LPX integrates 256 LP30 chips, 128GB of SRAM, and compiler-prescheduled inter-chip transfers into a low-latency decoding system. Third-party testing shows it leading existing public endpoints in long-context speed, but cost, concurrent throughput, and production-service performance remain undisclosed.

2026-08-25