RMIT-ADM+S 团队在 TREC 2026 RAG 赛道：研究智能体、检索消融与开源模型复现
RMIT-ADM+S at TREC RAG 2026: Research Agents, Retriever Ablations, and Open-Weight Replication
搞 RAG 和智能体的同学可以看看这份 TREC 2026 参赛报告。七个方案横跨 OpenAI 和开源模型，消融结果指出写作模型选择比检索策略更关键，开源的 owa-glm5 甚至在人工评审里赢了一个批次。想用低成本模型搭研究智能体的话，这些数据很参考价值。
RMIT-ADM+S 团队参加 TREC 2026 RAG 赛道（使用 ClimbMix 语料库），提交了七个研究智能体运行方案，分别采用 OpenAI 专有模型与开源权重模型，测试不同的检索策略、写作模型与智能体工作流。提交前的消融实验表明，写作模型选择和相邻分块扩展是影响答案质量的两个主要因素。
在 NIST 官方评估中，七个运行方案中的六个在三套 rubric/nugget 集合上均超过赛道中位数。表现最好的 brief-hybrid 方案在其提交中取得最高的全项目覆盖率，并在 Batch-1 vital 覆盖率上以 0.892 追平赛道最高分。
人工偏好评估与自动覆盖率指标出现分歧：开源权重的 owa-glm5 尽管自动覆盖率较低，仍在 Batch 1 的人工对比中领先，而 sol-research-first 在 Batch 2 领先。这一结果说明自动覆盖指标与人类偏好之间存在差异，也显示低成本开源权重模型在研究智能体系统中具有实用潜力。
RMIT-ADM+S at TREC RAG 2026: Research Agents, Retriever Ablations, and Open-Weight Replication
We describe the RMIT-ADM+S team's participation in the TREC 2026 Retrieval-Augmented Generation (RAG) track using the ClimbMix corpus. We submitted seven research-agent runs using proprietary OpenAI and open-weight models, exploring different retrieval strategies, writer models, and agent workflows. Pre-submission ablations identified writer selection and adjacent-chunk expansion as the main factors affecting answer quality. In the official NIST evaluation, six of seven runs exceeded the track median across all three rubric/nugget collections. Our best-performing run, brief-hybrid, achieved the highest all-item coverage among our submissions and tied the track maximum (0.892) for Batch-1 vital coverage. Human side-by-side preferences differed across evaluation batches: the open-weight owa-glm5 led Batch 1 despite lower automated coverage, while sol-research-first led Batch 2. These findings highlight differences between automated coverage measures and human preferences, as well as the potential of lower-cost open-weight models for research-agent systems.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include TREC, RAG, owa-glm5, OpenAI.
Primary source verification: published directly by arXiv: OpenAI with first-party authority. Cross-source corroboration: 11 independent sources have covered this development (including Stanford HAI, TestingCatalog, Artificial Analysis), confirming industry-wide signal strength.