BestHub
Sign in
Tagged articles

Agent Training

13 articles · Page 1 of 1
SmartApe AI
SmartApe AI
Nov 5, 2026 · Artificial Intelligence

Harness-S: Internalizing Complementary Harness Experience for Generalizable Agents

Researchers from Shanghai Jiao Tong University and Xiaohongshu propose Harness-S, a framework that evolves complementary execution harnesses, uses them for selective rescue training on failed tasks, and adapts harness selection via a policy-aware router, enabling agents to internalize diverse problem-solving strategies and retain them in the base framework.

Agent TrainingBehavior TransferComplementary Learning
0 likes · 21 min read
Harness-S: Internalizing Complementary Harness Experience for Generalizable Agents
SmartApe AI
SmartApe AI
Oct 23, 2026 · Artificial Intelligence

IQuest-Q1 Technical Deep Dive: 320B MoE Agent Model with 512K Context

IQuest Research releases technical report for IQuest-Q1, a 320B MoE model with 15B active parameters, 512K context window, recursive multi-token prediction, and agent-focused post-training, achieving strong benchmarks across coding, tool use, and reasoning tasks.

Agent TrainingBenchmark EvaluationLong Context
0 likes · 16 min read
IQuest-Q1 Technical Deep Dive: 320B MoE Agent Model with 512K Context
SmartApe AI
SmartApe AI
Oct 16, 2026 · Artificial Intelligence

Meta's 816B Muse Spark Open Offers Full Agent Training Stack, But Chinese Models Lead

Meta released Muse Spark Open, an 816B parameter multimodal model for coding and long-horizon agents, accompanied by a complete post-training ecosystem including TorchTitan, SGLang, vLLM, Miles, and verl, with experimental results showing GRPO training improvements on math and geometry benchmarks, though Chinese models like Qwen, GLM, and Kimi already surpass it on coding benchmarks and have similar training infrastructure.

Agent TrainingChinese AI ModelsGRPO
0 likes · 8 min read
Meta's 816B Muse Spark Open Offers Full Agent Training Stack, But Chinese Models Lead
Xiaohongshu Tech REDtech
Xiaohongshu Tech REDtech
Oct 15, 2026 · Artificial Intelligence

CompoWorld: Compositional Environment Scaling Boosts Agent Performance by 9.17 Points

Xiaohongshu's AllSpark team introduces CompoWorld, a compositional environment scaling framework that combines 448 reusable services with 10,130 tools to train agents on cross-application workflows, achieving a 9.17-point average improvement across eight benchmarks and tripling success rates on AutomationBench.

Agent TrainingAllSparkAutomationBench
0 likes · 12 min read
CompoWorld: Compositional Environment Scaling Boosts Agent Performance by 9.17 Points
Machine Learning Algorithms & Natural Language Processing
Machine Learning Algorithms & Natural Language Processing
Oct 3, 2026 · Artificial Intelligence

DeepSeek DSec: Running 380K Sandboxes with 50x Overcommit for Agent Training

DeepSeek's DSec infrastructure supports millions of agent sandboxes through layered environments, on-demand image loading via 3FS, memory sharing with virtio-pmem/DAX, CPU scheduling, and trajectory forking, achieving 50x resource overcommit while addressing security challenges like agent-discovered vulnerabilities.

3FSAgent TrainingAppArmor
0 likes · 12 min read
DeepSeek DSec: Running 380K Sandboxes with 50x Overcommit for Agent Training
Old Zhang's AI Learning
Old Zhang's AI Learning
Jul 10, 2026 · Artificial Intelligence

NVIDIA Opens 10 Trillion‑Token Dataset to Power AI Agents

NVIDIA has open‑sourced a 10‑trillion‑token training corpus—including the Nemotron‑CC‑v2, Nemotron‑CC‑Math, and 53 million synthetic personas—paired with the Apache‑2.0 NeMo Data Designer pipeline, benchmarked improvements on math and code tasks, and tools for visualizing and generating data for AI agents.

Agent TrainingLarge Language ModelsNVIDIA
0 likes · 13 min read
NVIDIA Opens 10 Trillion‑Token Dataset to Power AI Agents
AI Architecture Hub
AI Architecture Hub
Jun 30, 2026 · Artificial Intelligence

How to Fine‑Tune LLMs in 2026: Overcome the 30‑40% Error Wall with GRPO and RULER

Teams building LLM‑powered products often hit a wall where 30‑40% of responses are wrong and the model never learns from mistakes; the article explains how modern fine‑tuning using GRPO‑based reinforcement learning and the open‑source ART framework, together with the RULER reward‑free evaluator, lets small open‑source models surpass larger ones in cost, latency, and accuracy.

ART frameworkAgent TrainingGRPO
0 likes · 9 min read
How to Fine‑Tune LLMs in 2026: Overcome the 30‑40% Error Wall with GRPO and RULER
Data Party THU
Data Party THU
Jun 14, 2026 · Artificial Intelligence

Stop Misunderstanding AI Agents: A Clear Guide to All Core Concepts

The article defines AI Agent as a system centered on a large model that can invoke tools, receive feedback, and continuously accomplish tasks, and systematically distinguishes related terms such as Model, Scaffolding, Harness, Context Engineering, Policy, Tool, Skill, Sub‑agent, Environment, Rollout, Reward, and Trainer, using concrete examples to clarify each.

AI agentAgent TrainingHarness
0 likes · 10 min read
Stop Misunderstanding AI Agents: A Clear Guide to All Core Concepts
Machine Learning Algorithms & Natural Language Processing
Machine Learning Algorithms & Natural Language Processing
May 30, 2026 · Artificial Intelligence

Breaking the Agent Training Bottleneck: Open‑Source ClawGym Data, Training, and Evaluation Pipeline

ClawGym provides a complete open‑source framework for Claw‑style personal agents, linking a 13.5 K synthetic task dataset, black‑box rollout training, sandbox‑parallel reinforcement learning, and a rigorously verified benchmark of 200 tasks, and demonstrates that synthetic data can lift a 30 B model beyond a 235 B baseline.

Agent TrainingClawGymOpenClaw
0 likes · 16 min read
Breaking the Agent Training Bottleneck: Open‑Source ClawGym Data, Training, and Evaluation Pipeline
Machine Learning Algorithms & Natural Language Processing
Machine Learning Algorithms & Natural Language Processing
Apr 25, 2026 · Artificial Intelligence

Why DeepSeek‑V4 Took Twice as Long: Inside the Training‑Stability Challenges and Engineering Hacks

The DeepSeek‑V4 technical report reveals that the model’s doubled training time stems from massive token and parameter scaling, severe training‑stability issues in MoE layers, and a suite of engineering solutions—including Anticipatory Routing, SwiGLU Clamping, specialist expert training, and a custom sandbox cluster—while also exposing high hallucination rates despite impressive benchmark performance.

Agent TrainingDeepSeek-V4Generative Reward Model
0 likes · 12 min read
Why DeepSeek‑V4 Took Twice as Long: Inside the Training‑Stability Challenges and Engineering Hacks
Baobao Algorithm Notes
Baobao Algorithm Notes
Apr 14, 2026 · Industry Insights

Why Mastering AI Agents Is the Most Critical Skill Right Now

The article argues that leveraging AI agents like Claude Code is now the top priority for developers, explaining how agents boost productivity, the importance of their operating environment, and why embracing them is essential for future success in the AI-driven workplace.

Agent TrainingClaude CodeEnvironment
0 likes · 10 min read
Why Mastering AI Agents Is the Most Critical Skill Right Now
Meituan Technology Team
Meituan Technology Team
Jan 29, 2026 · Artificial Intelligence

How LongCat‑Flash‑Thinking‑2601 Achieves Real‑World Generalization for Agents

LongCat‑Flash‑Thinking‑2601, a 560‑billion‑parameter MoE model, combines environment expansion, multi‑environment RL, systematic noise training, a heavy‑thinking reasoning mode, and Zigzag sparse attention to deliver strong benchmark performance and robust real‑world agent capabilities.

Agent TrainingEnvironment ExpansionReinforcement Learning
0 likes · 14 min read
How LongCat‑Flash‑Thinking‑2601 Achieves Real‑World Generalization for Agents